Where Stores Actually Lose the Sale: Ecommerce Conversion Funnel Drop-Off at Every Stage

An ecommerce conversion funnel has four transitions that a Shopify merchant can actually measure: session to product view, product view to add to cart, add to cart to checkout started, and checkout started to purchase. Of those four, two have credible published benchmarks, one has two incompatible ones, and one has none at all.

This post prints the numbers that exist, names the source and the date inside every cell, and leaves the fourth row empty on purpose. An empty cell is a finding. A filled-in cell with no traceable origin is a liability, and there are a great many of those on the first page of Google for this exact query.

The four transitions worth measuring, and the two vanity metrics to ignore

TL;DR A funnel stage is a transition between two steps, not a page. Four transitions matter. Traffic volume and bounce rate are not among them, because neither tells you which step lost the sale.

The word “funnel” gets used for two entirely different things, and the confusion costs merchants money.

The first thing is the marketing funnel: awareness, consideration, intent, purchase, loyalty. It is a model of a shopper's head. Nobody can measure it directly, which is why articles about it are full of hypotheticals. Shopify's own page on this topic, dated March 4, 2023, illustrates its funnel with exactly this kind of thought experiment, asking you to imagine that one million people have heard of your brand and 500,000 have considered buying (Shopify blog, fetched July 30, 2026). Useful for strategy. Useless for diagnosis, because you cannot open a report tomorrow and see how many people “considered.”

The second thing is the behavioural funnel: the sequence of things a browser actually did on your store, each one recorded. That is what this post is about, and it is the only version you can fix.

Google Analytics 4 defines it in five events. Its purchase journey report tracks session start (session_start), view product (view_item), add to cart (add_to_cart), begin checkout (begin_checkout) and purchase (Google Analytics Help, fetched July 30, 2026). Shopify Analytics reports three of the same boundaries under different names, which we will get to.

Five events means four transitions. Those transitions are the funnel:

1. Session → product view. Did the visitor find something to look at? This is your homepage, collection pages, search, and navigation doing their job or failing to.
2. Product view → add to cart. Did the product page persuade? This is imagery, price, reviews, variant selection, stock messaging, delivery promise.
3. Add to cart → checkout started. Did the cart survive contact with the total? This is where shipping cost, taxes, and thresholds do their damage.
4. Checkout started → purchase. Did the form and the payment step let them through? This is fields, errors, account walls, and payment method coverage.

Each transition has a distinct owner, a distinct set of fixes, and a distinct benchmark situation. That is the entire argument for measuring transitions rather than a single blended conversion rate. A store at 1.8% overall could be at 1.8% because its product pages are weak, or because its checkout is broken, or because its collection pages bury everything below the fold. The blended number cannot tell those apart. The four transitions can, in about ten minutes.

The two metrics to stop reporting

Sessions. Traffic volume is an input, not a funnel stage. It belongs in your acquisition review. Putting it at the top of a funnel chart invites the conclusion that the fix is more traffic, which is the most expensive fix available and usually the wrong one. Contentsquare's 2026 benchmark, drawn from 99 billion sessions across 6,500+ websites in the period Q4 2024 to Q4 2025, recorded conversion rates down 5.1% year over year across its whole 9-industry panel, not retail alone (Contentsquare). More traffic into a leakier funnel is a treadmill.

Bounce rate. Bounce is a symptom that spans two stages at once and localises neither. Contentsquare's January 28, 2025 press release, covering 90 billion sessions across 6,000 websites, reported that 53% of users exit after viewing just a single page (Contentsquare, data Q4 2023 vs Q4 2024). That figure is real and it is alarming, but it does not tell you whether those single-page exits were people who landed on a collection page and found nothing, or people who landed on a product page and were put off by the price. Different problem, different fix, same bounce rate.

Replace both with the four transition rates. Then you know where to look.

The reference table: drop-off at every stage, with the source and date in every cell

TL;DR Two transitions have solid figures, one has two incompatible figures we print side by side without averaging, and one has nothing. The confidence column is the most useful column in the table.

Here is the whole funnel in one place. Read the confidence column before you read the numbers.

Transition	Published benchmark	Denominator, as the source defines it	Source · date	Confidence
Session → product view	No figure printed	–	–	No credible public benchmark. Google's GA4 documentation defines the view_item stage and publishes no benchmark figures at all (fetched Jul 30, 2026). Shopify Analytics has no session-to-product-view report. Every circulating figure we traced led to a composite page with no sample size. We are not printing one.
Product view → add to cart	6.07% and 4.6%, printed side by side, never averaged	Dynamic Yield: “% of items added to cart after product page view(s) by visitors over the past twelve months.” Littledata: not disclosed on the page	Dynamic Yield, live page read Jul 30, 2026 · Littledata, 2,800 sites, 2023 study	Medium. Both vendor-published. The Dynamic Yield page is a rolling 12-month average that moves month to month. Littledata's is three years old and its denominator is undefined.
Add to cart → checkout started	About 75% reach checkout (the source states a 25% drop between add to cart and checkout)	Shopify online-store sessions. “Reached checkout” inherits Shopify's definition: sessions with user input during checkout	DTC Pages, Q2 2026 · 21 Shopify stores, 179M+ sessions · published Apr 4, 2026	Low to medium. n = 21 stores. It is the only transparent, current, session-level figure we could verify for this transition. Printing it with its n is the point.
Checkout started → purchase	48.4% complete (DTC Pages) and 45% complete (Littledata)	DTC Pages: Shopify sessions. Littledata: not disclosed on the page, which publishes the rate without defining what it divides by	DTC Pages, Q2 2026 · Littledata, 2,800 sites, 2023 study	Medium. Two independent panels landing within 3.4 points of each other. The best-corroborated transition in the table.
Whole funnel: session → purchase	2.16% mean, 2.07% median, 2.59% at the 75th percentile	Orders divided by online-store sessions per store, then averaged across stores	DTC Pages, Q2 2026 · same 21-store panel	Low to medium. Small panel, DTC and US skewed. For a wider read, see our overall Shopify conversion rate benchmarks.

All sources fetched or read on July 30, 2026. DTC Pages states its methodology verbatim: “Conversion rates come from Shopify's sessions table (online store sessions only), not total orders divided by total sessions.” Its panel is 21 Shopify stores representing $417M combined revenue and 179M+ sessions, covering January 2025 to June 2026. Dynamic Yield's figures are rolling 12-month averages that change; the linked page carries the live value.

Three observations about that table matter more than any single number in it.

First, the panel sizes are wildly uneven and you should care. The add-to-cart row rests partly on a panel that Dynamic Yield describes as aggregating over 200 million monthly unique users from more than 400 brands across 300 million-plus sessions. The add-to-cart-to-checkout row rests on 21 stores. Those are not the same class of evidence, and no honest table would present them in the same typeface without saying so. We say so.

Second, the two rows with the most solid evidence are the two closest to the money. Checkout completion is corroborated by two independent panels three years apart. That is not a coincidence: checkout is instrumented by everyone, because everyone is trying to sell you checkout software. The top of the funnel is instrumented by nobody, because there is no product to sell against it.

Third, the empty row is the most honest thing on this page. Search for a session-to-product-view benchmark and you will find plenty of numbers. We went looking for their origins. The next section is what we found.

The stages nobody has published a credible benchmark for, and why we are not printing one

TL;DR The session-to-product-view rate has no traceable public benchmark. Google defines the stage and publishes no number. Shopify does not report it. The figures in circulation trace to pages with no sample size, no field date, and no primary source.

This is the section that a benchmark post is not supposed to have, and it is the reason this one exists.

The first transition in every funnel diagram ever drawn is the one from “arrived” to “looked at a product.” It is arguably the most actionable stage a merchant has, because collection pages, search, and navigation are cheap to change. And there is no benchmark for it that we could verify.

Here is exactly what we checked and what each check returned.

• Google's official GA4 documentation. It defines the purchase-journey stages in detail and even works an abandonment example: “if 200 users begin a session and only 140 users view a product then the abandonment rate below the current step would be 60 (or 30%)” (Google Analytics Help, fetched July 30, 2026). Those are illustrative numbers inside a worked example. Google publishes no benchmark figures on that page. We fetched it to be sure.
• Shopify Analytics. Shopify's behaviour reports document three session metrics: sessions with cart additions, sessions that reached checkout, sessions that completed checkout (Shopify Help Center, fetched July 30, 2026). There is no sessions-that-viewed-a-product metric in that list. The stage is not in the product, so it is not in the benchmark.
• IRP Commerce's live market data. IRP publishes fresh monthly sector conversion rates from its own merchant network, and its June 2026 table runs from Arts and Crafts at 5.53% down to Baby and Child at 0.51% (IRP Commerce, UK data). We fetched it looking for stage data. IRP publishes no add-to-cart rate and no cart abandonment rate. That absence, from one of the freshest public datasets available, is itself informative.
• The figures in circulation. Every session-to-product-view number we chased led to a page that cited another page. None terminated at a study with a disclosed sample size and a field date. We document two of those chains in full further down.

So we leave the row empty. This is not modesty. It is the only defensible option, and it has a practical consequence for you: for this stage, your benchmark is your own trailing baseline. Measure your session-to-product-view rate in GA4, write it down, change one thing on your collection pages, and compare against yourself. There is no industry row to hide behind, which as it happens is how all four stages should be treated anyway.

The second gap: nobody publishes stage benchmarks by vertical that we could verify

The overall conversion rate splits enormously by category. Dynamic Yield's live benchmark hub shows conversion rates running from Beauty and Personal Care at 5.37% down to Luxury and Jewelry at 0.71%, with average order values inverted across the same span, from $63 in Pet Care to $386 in Luxury (Dynamic Yield benchmark hub, rolling 12 months, read July 30, 2026). We cover that split properly in our post on conversion rate by industry.

Two of the four transitions do have partial vertical cuts, and they are worth having. Dynamic Yield's add-to-cart page shows Food and Beverage at 10.19% against Luxury and Jewelry at 1.72%, a near six-fold span on the same metric (read July 30, 2026). Littledata's older study gives fashion 5.4% and food and beverage 4.8% (2,800 sites, 2023). Its cart-abandonment page splits by vertical too, with Beauty at 79.81% and Pet Care at 51.1% (Dynamic Yield, read July 30, 2026).

What does not exist, as far as we could verify, is a per-vertical breakdown of checkout completion or of add to cart to checkout. If your category is unusual, and jewellery and furniture both are, you are again benchmarking against yourself on those two rows. Say so out loud in your reporting rather than quietly borrowing the cross-industry figure.

Why five published add-to-cart rates do not reconcile

TL;DR Four publishers use the phrase “add-to-cart rate” for four different fractions. Dynamic Yield divides by product views. DTC Pages divides by sessions. Littledata does not say. Shopify counts sessions, GA4 counts users. Comparing across them is meaningless.

If you have ever put your add-to-cart rate next to a published benchmark and felt confused, this is why.

What it is called	Who publishes it	The actual denominator	Value
Add-to-cart rate	Dynamic Yield	Items added to cart, divided by product page views, over the past twelve months (verbatim on page)	6.07%
Add-to-cart rate	Littledata	Not disclosed on the page	4.6%
Session to add to cart	DTC Pages	Shopify online-store sessions with a cart addition	5.95%
Sessions with cart additions	Shopify Help Center (official)	“the total number of sessions where a customer added a product to a cart”	your number
Add to cart	Google Analytics 4 (official)	Users who fired add_to_cart, inside a closed funnel	your number

All five fetched or read July 30, 2026. Dynamic Yield's is a rolling 12-month average and changes. Littledata's is a 2023 study across 2,800 sites. DTC Pages covers 21 Shopify stores, 179M+ sessions, Jan 2025 to Jun 2026.

Look at what happens if you ignore the fourth column. A store measuring sessions with cart additions reads 4.2%, compares itself to Dynamic Yield's 6.07%, and concludes the product pages are failing. But Dynamic Yield's denominator is product page views, and a single session frequently contains several. If your average product-viewing session sees two products, your session-based rate and your view-based rate can differ by roughly a factor of two before any real difference in shopper behaviour exists at all.

The gap runs the other direction too. Littledata's page reports 4.6% but never says what it divided by, which means nobody can compare against it correctly, including Littledata's own readers. That is not a criticism of the number. It is a criticism of publishing a rate without its fraction.

The one that will actually bite you: Shopify counts sessions, GA4 counts users

This is the most consequential thing in this post for a merchant who runs both tools, and it is fetch-verified from both official docs.

Shopify counts sessions. All three of its funnel metrics are session metrics. From the help page, verbatim: sessions with cart additions is “the total number of sessions where a customer added a product to a cart”; sessions that reached checkout is “the total number of sessions where there was user input (for example, a key press or mouse click) during checkout”; sessions that completed checkout is “the total number of sessions where a customer purchased a product” (Shopify Help Center, fetched July 30, 2026).

GA4 counts users. Its purchase journey report is user-counted and, critically, closed: “users are only counted in the steps they complete in the specified sequence” (Google Analytics Help, fetched July 30, 2026). A shopper who adds to cart on Monday and checks out on Thursday is two sessions in Shopify. In GA4's closed funnel they may not be counted as progressing at all, depending on how the sequence is configured.

And there is a third wrinkle inside the Shopify definition itself. “Reached checkout” does not mean arrived at checkout. It means there was user input during checkout. A shopper who lands on the checkout page, reads the shipping total, and closes the tab without typing anything may never appear in that metric. Your Shopify checkout-completion rate is therefore measured against a slightly self-selected group of people who at least started typing, which makes it flattering compared to a naive arrivals-based measurement.

None of this makes either tool wrong. It makes them non-interchangeable. Pick one as your system of record for funnel work, note which one in your reporting, and never mix a Shopify numerator with a GA4 denominator or benchmark one against the other.

Cart abandonment vs checkout abandonment: two different numbers, constantly quoted as one

TL;DR Cart abandonment is roughly 70%. Checkout abandonment is roughly 50%. Both are correct, because they divide by different things. Benchmarking your checkout number against the cart number will convince you of a crisis you do not have.

This is the single most common measurement error we see, and it is expensive because it sends merchants to fix the wrong stage.

Claim	Denominator	Value	Source · date
Cart abandonment	Carts created that did not become orders, meta-averaged across 50 published studies	70.22%	Baymard Institute, last updated Sep 22, 2025
Cart abandonment	Same construct, Baymard's other live page	70.19%	Baymard Institute, read Jul 30, 2026
Cart abandonment	Live vendor panel, 400+ brands	77.54%	Dynamic Yield, read Jul 30, 2026
Checkout abandonment	Sessions that reached checkout and did not buy	51.6%	DTC Pages, Q2 2026, 21 Shopify stores
Checkout abandonment	Not disclosed by the source; the page publishes the completion rate without its denominator	55% (100 minus 45% completion)	Littledata, 2,800 sites, 2023 study

Baymard states verbatim on the first page: “Based on the data we collected, we've calculated the average cart abandonment rate of 70.22%” and “This value is an average calculated based on 50 different studies.” On its checkout-usability page, read the same day, it states: “At Baymard we've tracked the global average cart abandonment rate for 14 years, and it currently sits at 70.19%.” We verified both values on both live pages on July 30, 2026 and report both rather than picking.

The arithmetic explains itself once the denominators are visible. Cart abandonment starts counting the moment an item enters a cart. Checkout abandonment only starts counting once a shopper has already committed enough to open checkout, which is a much more motivated group. Naturally a higher share of them buys. 70% and 52% are both true, and they are not in conflict.

Now watch what happens when the two get merged, which is what a great many blog posts do. A Shopify store pulls its own checkout-completion number, finds 55% of checkout sessions leaving, reads that “the average cart abandonment rate is 70%”, and concludes it is doing fine. It is not doing fine or badly. It has compared a number to an unrelated number and learned nothing. Run it the other way and the store panics about a checkout that is performing at the panel median.

Baymard publishes two values for the same metric, on the same day

We want to be careful here, because this is easy to read as a gotcha and it is not one. Baymard's cart-abandonment figure is a running meta-average across 50 studies conducted between 2006 and 2025, and the individual studies on that page range from 55.00% (Forrester Research, 2010) to 84.27% (SaleCycle, 2020). When you average 50 heterogeneous studies and update the set over time, two pages of the same site can easily carry values from slightly different snapshots. A three-hundredths-of-a-point discrepancy is a rounding artefact of an honest methodology, not an error.

We flag it for one reason: it demonstrates that even the most careful publisher in this space produces a moving number, and any post quoting “70.22%” without a retrieval date is quoting a photograph as if it were a constant. If you use that figure in a board deck, use it with the date attached.

The 77.54% from Dynamic Yield is a third thing again: a live vendor panel rather than a meta-analysis, on a rolling twelve-month window, with regional splits running from Americas at 74.46% to APAC at 81.21% and device splits from desktop at 69.19% to mobile at 79.92% (read July 30, 2026). Do not average it with Baymard's. Different construction, different population. If mobile is the bulk of your traffic, note that this panel puts mobile cart abandonment more than ten points above desktop, which is the clearest single argument we can point at for treating mobile as its own funnel. The tactical side of that stage lives in our guide to reduce cart abandonment on Shopify.

Pulling each number out of Shopify Analytics, click by click

TL;DR Shopify gives you three of the four transitions out of the box, under the names sessions with cart additions, sessions that reached checkout, and sessions that completed checkout. It does not give you session to product view. Historical session data starts October 1, 2022.

You do not need a new tool for this. You need four numbers from a report you already have.

In your Shopify admin, go to Analytics, then Reports, then the behaviour reports. The three metrics you want, with Shopify's own definitions, are:

1. Sessions with cart additions. Shopify defines it as “the total number of sessions where a customer added a product to a cart.” Divide by total online-store sessions for the same window to get your session-to-add-to-cart rate. This is the number comparable to DTC Pages' 5.95%, and to nothing else in this post.
2. Sessions that reached checkout. Shopify defines it as “the total number of sessions where there was user input (for example, a key press or mouse click) during checkout.” Divide by sessions with cart additions for your add-to-cart-to-checkout rate. Compare to about 75%.
3. Sessions that completed checkout. Shopify defines it as “the total number of sessions where a customer purchased a product.” Divide by sessions that reached checkout for your checkout completion rate. Compare to 48.4% and 45%.
4. Conversion rate over time. Shopify defines this as “the percentage of online store visitors that make a purchase over a selected period of time.” This is your whole-funnel number, and it should be close to the product of the three transition rates above.

All four definitions are verbatim from the Shopify Help Center, fetched July 30, 2026.

Three traps in that report

Trap one: the data has a hard floor. Shopify's help page states that these session metrics have historical data only from October 1, 2022 onward. If you are building a multi-year trend, that is where it starts, and any earlier comparison you make is either a different metric or an estimate.

Trap two: “reached checkout” is not “arrived at checkout.” It requires user input. This means your checkout-completion denominator excludes an unknown number of people who opened checkout, saw the total, and left without touching the keyboard. Those shoppers are real abandonment, and Shopify's checkout metric is structurally blind to them. If you want to catch that group, you need the transition measured from cart-page views or from a begin_checkout-style event, which is exactly why the Shopify and GA4 numbers will never match.

Trap three: it will not equal GA4, and that is expected. Sessions versus users, open versus closed funnel, input-based versus event-based checkout entry. Pick your system of record and stay in it. A funnel measured two ways is a funnel you cannot act on.

What Shopify does not give you, and how to get it

There is no session-to-product-view metric in Shopify's behaviour reports. For that first transition you need GA4's purchase journey report, where the stage exists as view_item, or an equivalent event stream. Set it up if you can, because that stage is often the cheapest one to fix. Just remember, per the row we left blank, that once you have the number there is nothing credible to compare it against except your own history.

Finding your worst stage in ten minutes

TL;DR Pull four numbers, compute three ratios, compare each to its benchmark, and rank by the size of the gap. Do not fix the stage with the biggest drop-off. Fix the stage with the biggest gap versus its own benchmark.

The most common mistake at this point is to look at the funnel, see that the biggest absolute drop is at the top, and go fix the top.

Every ecommerce funnel loses most of its people at the first two stages. That is not a defect, it is what browsing is. Roughly 94 out of every 100 sessions will not add anything to a cart even on a perfectly healthy store, because DTC Pages' panel-wide session-to-add-to-cart figure is 5.95%. The size of a drop tells you nothing. The gap between your drop and the benchmark drop tells you everything.

Here is the ten-minute version.

1. Pick a clean window. A full calendar month, no Black Friday, no site migration, no sitewide sale. Mixed-promotion months produce funnels that describe the promotion rather than the store.
2. Pull four numbers. Total online-store sessions, sessions with cart additions, sessions that reached checkout, sessions that completed checkout.
3. Compute three rates. Cart additions divided by sessions. Reached checkout divided by cart additions. Completed divided by reached.
4. Write each next to its benchmark. 5.95%, about 75%, and 48.4% respectively, all from the 21-store DTC Pages panel for Q2 2026. Use the same panel for all three so at least your comparison is internally consistent, then cross-check the third against Littledata's 45%.
5. Express each gap as a ratio, not a difference. If your add-to-cart rate is 3.8% against a 5.95% benchmark, that is 64% of benchmark. If your checkout completion is 44% against 48.4%, that is 91% of benchmark. The first gap is far bigger, even though the second one is bigger in percentage points.
6. Split by device before you conclude anything. The same panel reports mobile conversion at 2.29% on 85.9% of traffic against desktop at 3.74% on 12% of traffic (DTC Pages, Q2 2026). A blended funnel with that traffic mix is basically a mobile funnel wearing a disguise.
7. Rank by dollars, not by gap. Which is the next section.

One caution about step 5. All three benchmarks come from a 21-store panel. If your first reaction to a gap is “we are 36% below benchmark,” your second reaction should be “below a benchmark drawn from 21 DTC stores, which may or may not resemble mine.” Use the ranking to decide what to look at first. Do not use it to decide what is broken. Looking at the actual pages is what decides that, and the four diagnostic sections below are about exactly that.

The formula: turning a stage drop-off into dollars at risk this month

TL;DR Sessions, times the gap between your stage rate and the benchmark, times your own downstream rates, times AOV, times contribution margin. The result is a ceiling for prioritising, not a forecast. Incremental carts convert worse than existing ones.

A funnel gap in percentage points does not compete for attention against anything else on your roadmap. A funnel gap in dollars does.

The formula is one line:

Monthly revenue at risk = Sessions × (benchmark stage rate − your stage rate) × (your own downstream stage rates, multiplied) × your AOV × your contribution margin

Every benchmark input below is cited. Every store input is labelled as yours to replace.

Input	Value	Where it comes from
Monthly online-store sessions	50,000	Your Shopify Analytics
Your session to add-to-cart rate	3.80%	Your sessions with cart additions, divided by sessions
Benchmark for that stage	5.95%	DTC Pages, 21 Shopify stores, 179M+ sessions, Q2 2026
Gap	2.15 points → 1,075 extra carts	Derived: 50,000 × 2.15%
Your add-to-cart to checkout rate	75.0%	Yours. Shown here at the DTC Pages value as a placeholder
Your checkout to purchase rate	48.4%	Yours. Shown here at the DTC Pages value as a placeholder
Survival through both	36.3% → 390 extra orders	Derived: 0.750 × 0.484 × 1,075
Average order value	$74.12	Triple Whale median AOV, its own brand panel, paid channels, 2025. Replace with your own.
Gross revenue at risk	about $28,900 per month	Derived: 390 × $74.12
Contribution margin	Yours	What the order is actually worth after COGS, shipping, payment fees and returns

Benchmark inputs cited; store inputs are placeholders. The AOV figure is the median across Triple Whale's stated panel of over 33,000 brands representing $18.4B in ad spend, on paid channels, published on a page last updated February 26, 2026. Use your own AOV: a median across 33,000 brands is not your store.

The gross number is the one people quote in meetings. The number that should drive the decision is the one after contribution margin, because a $28,900 revenue gap on a 22% margin is a $6,400 profit gap, and that is what competes against the cost of the work. We break the calculation down properly in our guide to what a recovered order is actually worth.

The ceiling caveat, which is not optional

This calculation produces an upper bound, not a forecast, and there is a specific structural reason why.

The formula assumes that incremental carts convert downstream exactly like existing ones. They do not. Visitors who needed persuading to add to cart are, by construction, less committed than visitors who added unprompted. Push your add-to-cart rate from 3.8% to 5.95% and the marginal 1,075 carts will contain a higher share of window-shoppers, price-checkers and people comparing you against three other tabs. Their checkout completion will run below your existing 48.4%, sometimes well below.

There is published evidence for treating upstream metrics with this suspicion. Baymard's cart-abandonment research found that 42% of US online shoppers who abandoned a cart did so because they were “just browsing / not ready to buy” (Baymard, fetched July 30, 2026; the page describes this as its latest quantitative study but discloses no sample size and no field date, which we note because we are holding ourselves to the same standard we apply to everyone else). Nearly half of all cart abandonment is not a problem with your store. It is people using a cart as a shortlist. Adding more of those people to your carts adds very little revenue.

So use the dollar figure the way it is meant to be used: as a ranking across your four stages, not as a promise. Compute all four, sort descending, work the top one. Then run the change as a real test and check whether the change is real before you bank it, because a stage rate that moves the week after a deploy has moved for many reasons, and only one of them is your deploy. If the test looks flat, our post on A/B test sample size explains why that is usually a traffic problem rather than an answer.

Session to product view: when your collection and search pages are the leak

TL;DR No published benchmark exists for this stage, so measure yourself against yourself. The evidence that does exist points at session depth and at traffic quality, not at the product pages.

A visitor who never sees a product page cannot buy anything, and the reasons they never saw one are almost always structural rather than persuasive.

Since we cannot hand you a benchmark for this transition, here is what we can hand you: the two things that are documented as moving it, and the diagnostic questions that localise the problem.

Session depth correlates with conversion

Contentsquare's 2025 benchmark analysis, across 90 billion sessions and 6,000 websites, found that sites which increased session depth by 10% or more saw “an average 5.4% boost in conversions” (Contentsquare press release, January 28, 2025). Read that as a correlation across 6,000 sites, because that is what it is. It is not a causal lift and Contentsquare does not claim it is. Deeper sessions and higher conversion travel together; a site that gets people to a second and third page is doing something right, but forcing extra pageviews will not manufacture the conversion.

The same dataset found 53% of users exit after viewing just a single page. If you sell more than one thing, that number is the size of your discovery problem.

Traffic source predicts this stage more than design does

Contentsquare's conversion guide, drawn from 99 billion sessions across 6,000+ sites for Q4 2024 to Q4 2025, reports conversion by channel: paid search converts at 2.8%, organic social at 0.7%, a four-fold spread (Contentsquare). The same source reports returning visitors converting at 2.9% against new visitors at 1.7%, with return visits making up 52.8% of traffic.

That press release also documented what happens when the mix shifts: sites whose traffic skewed heavily to paid social saw bounce rates up 9.2%, page views down 8.7% and conversions down 10.6% (Contentsquare, January 28, 2025). If your session-to-product-view rate fell last quarter, check your channel mix before you check your collection page layout. A shift toward cold paid social will move that stage on its own with nothing changed on the site.

The diagnostic questions for this stage

• Where do sessions land? If most land on the homepage, the homepage is a routing problem. If most land on collection pages, the collection page is a merchandising problem. If most land on product pages already, this stage barely exists for you and you should move down the funnel.
• How many products are above the fold on your top collection page, on a phone? Mobile is 85.9% of traffic in the DTC Pages panel and 69.9% in Contentsquare's much larger one. Whatever your figure, the mobile collection page is the real collection page.
• Does site search return results for your top ten queries? Zero-result searches are a discovery leak that never shows up as a bounce, because the shopper technically viewed a page.
• Is the first product view happening on the right product? A collection sorted by manual order that has not been touched in a year is showing new visitors last summer's priorities.

Product view to add to cart: when the product page is the leak

TL;DR Two published benchmarks exist and they measure different fractions: 6.07% of product views (Dynamic Yield, live) and 4.6% with an undisclosed denominator (Littledata, 2023). The vertical spread is nearly six-fold, so read your own category row.

This is the stage where a benchmark is most likely to mislead you, because the two available figures are not the same measurement.

Dynamic Yield's live benchmark reports a global add-to-cart rate of 6.07%, defined on the page as the “% of items added to cart after product page view(s) by visitors over the past twelve months” (read July 30, 2026, and the value moves as the rolling window advances). Littledata's study of 2,800 sites reports 4.6% with top-20% performance above 7.5% and top-10% above 9.6%, but does not define its denominator anywhere on the page (2023 study). DTC Pages reports 5.95% of Shopify sessions (21 stores, Q2 2026).

Three numbers, three different fractions. Match yours to whichever denominator you actually measured, and ignore the other two.

The cuts inside the add-to-cart benchmark that are worth having

Cut	Segment	Add-to-cart rate	Source
Global	Whole panel	6.07%	Dynamic Yield, read Jul 30, 2026
Region	EMEA	6.48%	Dynamic Yield
Region	Americas	6.1%	Dynamic Yield
Region	APAC	3.16%	Dynamic Yield
Device	Mobile	6.31%	Dynamic Yield
Device	Tablet	6.1%	Dynamic Yield
Device	Desktop	5.26%	Dynamic Yield
Vertical	Food & beverage (panel high)	10.19%	Dynamic Yield
Vertical	Luxury & jewelry (panel low)	1.72%	Dynamic Yield
Vertical	Fashion	5.4%	Littledata, 2,800 sites, 2023
Vertical	Food & beverage	4.8%	Littledata, 2,800 sites, 2023

Dynamic Yield rows are all on its product-view denominator and read on July 30, 2026; the page is live and the values move. Littledata rows are from a 2023 study with an undisclosed denominator, and are shown in the same table only to make the incompatibility visible. Note that the two panels disagree on direction for food and beverage, which is what non-comparable denominators look like.

Two things in that table deserve a second look.

Mobile adds to cart more often than desktop in this panel. 6.31% against 5.26%. That inverts the usual story, and it is consistent with adding to cart being a low-commitment action on a phone. It is also a reminder that the mobile conversion problem, which is real, lives further down the funnel rather than at this stage. Dynamic Yield's own device split puts mobile at 75.92% of sessions against desktop at 23.08% in the same panel.

The vertical spread is nearly six-fold within one panel. Food and beverage at 10.19% against luxury and jewelry at 1.72%. Consumables get added to carts constantly and cheaply. High-consideration goods do not. If you sell furniture and your add-to-cart rate is 2%, the global 6.07% figure is not your target and never was.

What actually moves this stage

We are going to be disciplined here, because this is the stage where vendor lift claims are thickest and almost none of them survive a source check. What we will say is what the funnel structure itself implies.

Product view to add to cart is a persuasion transition. Everything on the page that answers a shopper's open question moves it: whether the item will fit, what it looks like on something other than a white background, when it arrives, what happens if it is wrong, whether other people liked it, and what it actually costs including delivery. Every one of those is a hypothesis you can test on your own store, and the honest way to find out which matters for you is to test it rather than to import a case study from a store that is not yours. The mechanics of doing that on a Shopify product page are in our guide to Shopify product page optimization.

One structural note worth more than any tactic. If this stage is weak but your checkout completion is strong, you have a persuasion problem, and traffic quality is the first suspect. If this stage is strong and everything below it is weak, you have a friction problem, and the fixes are mechanical. The four transitions are diagnostic precisely because they separate those two cases.

Add to cart to checkout: when shipping cost is the leak

TL;DR About 75% of carts reach checkout in the one transparent panel we could verify (21 Shopify stores, Q2 2026). Extra costs are the top non-browsing reason shoppers abandon, at 40% in one Baymard study and 48% in a different one. Do not merge those two figures.

This is the shortest transition in the funnel and the one with the clearest single cause.

DTC Pages' panel puts the drop between add to cart and checkout at 25%, so roughly three in four carts make it to checkout (21 Shopify stores, 179M+ sessions, Q2 2026). We could not find a second independent, current, session-level figure for this specific transition anywhere, which is why this row carries a low-to-medium confidence rating in the master table. Take it as an order of magnitude.

What is much better documented is why people leave at this point. Baymard asked US online shoppers who abandoned carts for their reasons. The single largest answer, at 42%, was “I was just browsing / not ready to buy,” which is not a fixable store problem. Excluding that group, Baymard publishes this distribution:

Reason	Share	Which stage it hits
Extra costs too high (shipping, tax, fees)	40%	Cart to checkout
Delivery was too slow	20%	Cart to checkout
Didn't trust the site with credit card information	19%	Checkout to purchase
The site wanted me to create an account	18%	Checkout to purchase
Checkout was too long or complicated	17%	Checkout to purchase
Website had errors or crashed	17%	Any stage
Returns policy was not satisfactory	13%	Product view to cart
Couldn't see or calculate total cost upfront	12%	Cart to checkout
Credit card was declined	10%	Checkout to purchase
There weren't enough payment methods	9%	Checkout to purchase
Unknown	7%	–

Baymard Institute, fetched July 30, 2026. Baymard's verbatim framing: “However, if we ignore the 'just browsing' segment, and instead look at the remaining reasons for abandonments we get the following distribution.” The denominator is the non-browsing segment only. Percentages sum to more than 100% because respondents could select multiple reasons. Baymard publishes no sample size and no field date for this study on that page, which we flag rather than hide. The stage column is our own mapping, not Baymard's.

Add up the cost-related rows on that list. Extra costs at 40%, cost not visible upfront at 12%, delivery too slow at 20%. Whatever the overlap between them, the cluster is dominant, and every item in it becomes visible to the shopper at exactly the same moment: the moment the cart shows a total. That is why this transition is a distinct funnel stage and not just an extension of the product page.

Two Baymard studies say 40% and 48%. They are not the same study.

This is a trap worth naming, because the two figures circulate as if one updated the other.

eMarketer reported that 48% of US adults abandoned at checkout because extra costs were too high, naming Baymard Institute as the source and disclosing the sample as 1,012 US adults aged 18 and over, fielded February 7, 2024 (eMarketer, published June 7, 2024).

The 40% above comes from a different Baymard study, on a different base (the non-browsing segment), with no published sample size or field date. 48% is not an update to 40%. They are two studies with different populations, different bases and different years, and merging them or presenting either as “the” number is exactly the kind of quiet error this post exists to avoid. If you cite one, cite it with its scope attached.

The practical read is the same either way. Extra costs are the top fixable reason people abandon in both studies, by a wide margin, and the fix is not necessarily free shipping. It is early shipping information, so that the total is not a surprise at the moment of commitment. The full set of tactics for this stage is in our guide to reduce cart abandonment on Shopify.

Checkout to purchase: when payment and form friction are the leak

TL;DR Roughly half of checkout sessions do not buy: 48.4% completion in one panel, 45% in another. Baymard rates 65% of 335 benchmarked checkout flows as mediocre or worse, and the newest public figure on checkout form length is from 2019.

This is the best-measured stage in the funnel and, judged by the published UX research, the most consistently broken one.

Start with the rates. DTC Pages puts checkout completion at 48.4% across its 21-store Shopify panel for Q2 2026. Littledata, from an entirely different panel of 2,800 sites in a 2023 study, puts Shopify's average checkout completion rate at 45%. Its page states the rate and never defines the denominator, which is the same gap we flagged on its add-to-cart page, so the two figures are close but not strictly like for like. Two independent measurements, three years apart, 3.4 points apart. In a field where benchmarks routinely differ by a factor of two, that is close to consensus.

Littledata also publishes the distribution, which is more useful than the average:

• Average checkout completion: 45% (Shopify stores)
• Mobile 44%, desktop 49%
• Top 20% of stores: above 59%
• Top 10% of stores: above 66%

All four from Littledata, 2,800 ecommerce sites, 2023 study. The gap between the median store and the top decile is 21 percentage points on the same stage, which is a much larger prize than anything available at the top of the funnel, and it is available to a merchant who never buys another click.

How bad is the average checkout, actually?

Baymard has the only large-scale public audit of this we could verify. Its checkout usability research states that it “benchmarked the checkout flows of 335 top-grossing US and EU e-commerce sites across our 110+ Cart & Checkout guidelines,” with “30,000+ checkout elements manually reviewed” (Baymard, fetched July 30, 2026). Its verdict:

• 65% of sites have a checkout performance of “mediocre” or worse
• 35% rate “decent” or better
• 2% rate “good”
• The average site has 32 unique improvements available in its checkout flow

Baymard also publishes a 35% potential improvement figure for checkout conversion. We are labelling that clearly: it is Baymard's modeled potential, not a measured A/B result. Treat it as a statement about how much headroom the research thinks exists, not as a lift you should expect.

The form-field figure is seven years old, and that is the finding

Checkout form length is the single most-cited lever for this stage. Here is the freshest public number we could verify: the average ecommerce site has 12.8 form fields in its checkout flow, against an achievable 6 to 8, which was itself a 14% improvement on the 14.88 average measured in 2016 (Baymard, based on 7,900+ manually rated checkout UX scores for the world's 60 largest ecommerce sites).

That page was published June 21, 2019. It is seven years old. It predates Shop Pay's dominance, express wallet buttons on the cart page, and Shopify's checkout extensibility migration. We are quoting it because it is the newest credible public figure on the topic that we could find, and the fact that the newest credible public figure on checkout form length is from 2019 tells you something about the state of this literature. Use it as a target shape (halve your fields) rather than as a current benchmark.

Diagnosing your own checkout stage

Map the Baymard reason list onto the mechanics of your checkout and you get a short, specific audit:

• Count your fields. On mobile, in a real checkout, with a real product. If you are above 12, you are at or above the 2019 average of the largest sites in the world, which is not a compliment.
• Check for the account wall. 18% of non-browsing abandoners cited forced account creation (Baymard). Guest checkout is not a nice-to-have.
• Count your payment methods against your market. 9% cited insufficient payment methods, 10% cited a declined card (Baymard). The second one is partly a payments-configuration problem, not a UX one.
• Look for errors on real devices. 17% cited website errors or crashes (Baymard). A checkout that fails on one browser version is invisible in aggregate reporting and fatal to the sessions it hits.
• Split mobile from desktop. 44% against 49% completion in Littledata's panel (2023). If mobile is 70% to 86% of your traffic, per the panels above, that five-point gap is where most of your lost orders live.

One caution on the trust row. 19% of non-browsing abandoners said they did not trust the site with their credit card information. That is a real signal, but it is not measurable from your analytics, which means it is a hypothesis to test rather than a diagnosis to accept.

Where the numbers in circulation actually came from

TL;DR We traced the widely quoted funnel benchmarks back to their origins. One is a top-quintile threshold that became an average across two hops. A page badged 2026 is still serving 2018 data. And the most-used AI assistant answers this question from two domains that are weeks old.

Everything above is a claim about sourcing discipline. This section is the evidence for it, and every chain here is reproducible by anyone with a browser.

Chain A: a percentile became an average

Follow this one carefully, because it is the tidiest example of how a benchmark gets corrupted without anyone lying.

1. Origin. Littledata publishes an add-to-cart rate: average 4.6%, with above 7.5% defined as the top 20% of stores (Littledata, 2,800 sites, 2023 study). Fetched July 30, 2026.
2. Hop one. Blend Commerce, a Shopify agency, reports it as: “Shopify-specific data from Littledata found an average of 4.4% exceeding 7.6% puts you in the top 20%” (Blend Commerce, fetched July 30, 2026). No sample size, no date, no link to the original. The average has drifted 0.2 points and the threshold 0.1.
3. Hop two. Triple Whale's ecommerce benchmarks page, last updated February 26, 2026 and backed by a stated panel of over 33,000 brands representing $18.4B in ad spend, gives an add-to-cart rate of “approximately 7.5%” and cites Blend Commerce for it (Triple Whale, fetched July 30, 2026).

7.5% was Littledata's top-quintile threshold. Two hops later it is circulating as the industry average on a page with a large and reputable panel behind it. A merchant reading that page benchmarks against a number that four out of five stores do not reach, concludes their product pages are failing, and goes to work on a problem they may not have. That is what an untraceable benchmark costs.

This is also why you will not find “a good add-to-cart rate is 7.5%” anywhere in this post.

Chain B: the number that lost its author

Baymard computes cart abandonment at 70.22% across 50 studies and says so explicitly on the page. The same Triple Whale page attributes 70.22% to Statista, not to Baymard. We are describing what Triple Whale cites; we did not fetch Statista and are not treating the figure as a Statista number.

Meanwhile Store Growers publishes a page titled “40+ Ecommerce Metrics Benchmarks (2026)”, last updated 8 January 2026. Its add-to-cart rate is sourced from Monetate's Ecommerce Quarterly for Q2 2018 (desktop 10.93%, mobile 9.78%, tablet 12.20%). Its cart abandonment figure is Baymard 2020's 69.80%. Its conversion rate comes from a 2018 Compass report, and its pages-per-session from Wolfgang 2020 (Store Growers, fetched July 30, 2026).

We are citing those figures only as evidence of staleness, not as benchmarks. A 2018 add-to-cart rate on a page badged 2026 is the single clearest illustration available of why a date matters more than a decimal place.

Chain C: the AI layer, measured

On July 30, 2026 we asked ChatGPT, via the DataForSEO ChatGPT scraper, for stage-by-stage funnel drop-off benchmarks. It returned three separate numeric tables, opening with: “if you're looking for good cross-industry benchmarks for a typical DTC ecommerce store, the following ranges are widely representative.”

Funnel stage	ChatGPT's “typical conversion”	ChatGPT's “typical drop-off”
Session → product view	45–60%	40–55%
Product view → add to cart	8–15%	85–92%
Add to cart → checkout	40–60%	40–60%
Checkout → purchase	50–70%	30–50%

These are not benchmarks and we do not endorse a single figure in this table. It is reproduced as evidence of what the most-used AI assistant currently answers to this question, retrieved via the DataForSEO ChatGPT scraper on July 30, 2026. The provenance audit below is the reason for the disclaimer.

The answer also asserted that “these ranges align reasonably well across GA4 purchase funnels, Shopify ecosystem benchmarks, Littledata-derived datasets.” It gave exactly three citations. We checked all three.

• Citation one: Google's GA4 documentation. We fetched it. It publishes no benchmark figures at all. It defines the stages. It cannot support any of the ranges above.
• Citation two: a domain first seen in the DataForSEO backlink index on July 2, 2026. That is 28 days before the query. It carries 11 backlinks from 5 referring domains, a spam score of 25, and returns no rank in DataForSEO's 0 to 100 domain rank. It publishes stage-to-stage rates by vertical with no sample size, no data date, and no author. Its only sourcing line describes the figures as median funnel stage conversion rates by vertical for Shopify and WooCommerce DTC stores. That is the entire methodology disclosure.
• Citation three: a domain first seen on May 22, 2026. 69 days old at query time, 48 backlinks from 12 referring domains, 2 crawled pages, domain rank 20. To its credit it openly describes its own numbers as a composite “compiled from global industry research” naming Littledata, Shopify, IRP Commerce, Baymard and Klaviyo, and calls them “directional benchmarks” and “a starting point, not your number.” It publishes no sample size and no per-metric dates.

For scale, the same DataForSEO domain-rank call on the same day returned 74 for contentsquare.com, 60 for baymard.com, 47 for littledata.io, 43 for marketing.dynamicyield.com and 15 for dtcpages.com. All figures from DataForSEO backlinks endpoints, July 30, 2026.

Stated precisely: the most-used AI assistant, asked this exact question today, produced a confident four-stage benchmark table whose only two substantive citations are a 28-day-old domain with five referring domains and no disclosed methodology, and a 69-day-old two-page domain that openly describes its numbers as a composite of other people's research. Neither publishes a sample size. Its third citation publishes no benchmarks at all.

We are not printing those domains' figures anywhere in this article, including in the table above, which reproduces only what ChatGPT itself said. We are naming what happened, because the same vacuum that produced that answer is the reason a merchant cannot currently get a trustworthy answer to a basic question about their own funnel.

The organic results are not much fresher

We pulled the live desktop SERP for “ecommerce conversion funnel” in the US on July 30, 2026 via DataForSEO. After the AI Overview, the first organic result is an Adobe page dated November 18, 2022. Then heap.io with no date displayed, Yotpo dated April 7, 2025, Bloomreach dated June 17, 2024 on a URL whose slug still reads “in 2021”, FullStory with no date, a discussions block surfacing Quora threads from 2020 and 2015, and Shopify's own funnel page, dated March 4, 2023, at absolute position 10.

We fetched the two leaders to check their contents. Shopify's page contains zero stage benchmarks; its only numbers are the hypothetical about one million people hearing of your brand. The heap.io guide contains no stage benchmarks either: its only figures are an overall conversion rate of 2.86% attributed to Invespcro in 2020 and a category table from IRP Commerce dated 2019, on a page that displays no publication or update date at all.

That is the state of the answer for this question in July 2026. It is not that the published benchmarks are wrong. It is that for two of the four transitions there are barely any, and the pages ranking for the query are not the ones that have them.

Methodology, and how to check every number in this article yourself

TL;DR Every figure here was read off a page we fetched on July 30, 2026 or returned by a named tool call on that date. Where a source publishes two values, we print both. Where no credible figure exists, we print nothing.

This is the part that lets you audit us.

The rules we applied

1. Fetch, then quote. No figure appears in this post unless we loaded the source page and confirmed the number is on it, or unless it was returned by a tool call we made on July 30, 2026. Nothing is quoted from memory or from a secondary citation, with one deliberate exception: the Statista attribution in Chain B, which we describe as what Triple Whale cites and explicitly do not treat as a verified Statista figure.
2. Print the denominator. Where a source defines its metric, we quote the definition verbatim. Where it does not, as with Littledata's add-to-cart page, we say so in the cell.
3. Print the sample size. Including when it is unflattering. The 21-store panel behind three of our headline transitions is stated everywhere those figures appear, because a reader who does not know n cannot weigh the number.
4. Print the date, and mark what moves. Dynamic Yield's benchmark pages are rolling twelve-month averages that change as the window advances. Every figure from them carries “read July 30, 2026” and will drift after publication. The linked page always carries the live value.
5. Never average across denominators. 6.07% and 4.6% and 5.95% are three different fractions. Their mean is a number that measures nothing, so we do not compute it.
6. When a source disagrees with itself, print both values. Baymard's 70.22% and 70.19% both appear here, with both URLs and the fetch date.
7. When there is no credible figure, print nothing. The session-to-product-view row is empty. It will stay empty until somebody publishes a study with a sample size and a field date.

How to reproduce this yourself in under an hour

• Open every linked source and find the number. Each external link in this article goes to a page where the figure we quoted appears. If one has moved, that is the rolling-source caveat doing its job, not a fabrication.
• Check whether the source states a denominator. For Dynamic Yield and Littledata, this takes thirty seconds each and is the single most revealing test you can run on any benchmark page.
• Check the date on the page, then check the date on the data. They are frequently years apart. Store Growers' page is the worked example: updated January 2026, serving 2018 data.
• Follow one citation upstream. Pick any benchmark blog post and click through its source. If the source is another blog post, click again. Chain A took two hops to reach the original, and the number had changed meaning by then.

What we would fix in this article if better data existed

Three things, stated openly so you can weight the post accordingly.

The 21-store panel is too small to carry three of our four transitions. We used it because it is the only source we found that publishes session-level Shopify funnel transitions with a stated methodology and a current date. A larger transparent panel would replace it tomorrow.

Two of our sources are three years old. Littledata's 2023 study underpins the checkout corroboration and one of the two add-to-cart figures. Three years in ecommerce spans a checkout platform migration and the arrival of express wallets on most storefronts.

Two of our sources are vendors publishing about their own customers. Dynamic Yield and Littledata both sell software to the stores in their panels. That is not disqualifying, and both are more transparent than most, but a panel of a vendor's customers is not a random sample of ecommerce.

A disclosure, since this post has been holding everyone else to a sourcing standard. We are building StorePilot, an AI CRO agent for Shopify that watches the four transitions above on your own store, points at the leaking one, and runs the test honestly rather than declaring an early winner. It has not launched yet. The arithmetic in this post works identically whether or not you ever use it, and the empty row in the master table stays empty in our product too.

Questions merchants keep asking

What are the stages of an ecommerce conversion funnel?

Four transitions, not four pages: session to product view, product view to add to cart, add to cart to checkout started, and checkout started to purchase. Google Analytics 4 names the same five events in its purchase journey report (session_start, view_item, add_to_cart, begin_checkout, purchase), and Shopify Analytics reports three of the four as sessions with cart additions, sessions that reached checkout, and sessions that completed checkout (Shopify Help Center and Google Analytics Help, both fetched July 30, 2026).

What is a good add-to-cart rate?

There is no single answer, because four publishers use the phrase for four different fractions. Dynamic Yield reports 6.07% globally and defines it as items added to cart after product page views (live page, read July 30, 2026). Littledata reports 4.6% across 2,800 sites but does not disclose its denominator on the page (2023 study). DTC Pages reports 5.95% of Shopify online-store sessions (21 stores, Q2 2026). Compare like with like before you judge your own number.

What percentage of shoppers abandon at checkout?

Roughly half, and that is a different number from cart abandonment. DTC Pages measured 51.6% of Shopify sessions that reached checkout leaving without buying (21 stores, 179M+ sessions, Q2 2026). Littledata's older study of 2,800 sites put checkout completion at 45%, implying 55% abandonment (2023). Baymard's widely quoted 70.22% is cart abandonment, a different denominator entirely.

What is the difference between cart abandonment and checkout abandonment?

The denominator. Cart abandonment counts carts created that never became orders, and Baymard's meta-average of 50 studies puts it at 70.22% (last updated September 22, 2025). Checkout abandonment counts only the sessions that actually reached checkout and did not buy, which DTC Pages measured at 51.6% (Q2 2026). Both are correct. Benchmarking your checkout number against Baymard's cart number will convince you of a crisis you do not have.

What is the average drop-off from product view to add to cart?

The two published figures we could verify are not comparable to each other. Dynamic Yield reports a 6.07% add-to-cart rate defined as items added after product page views (read July 30, 2026, and the value moves). Littledata reports 4.6% but publishes no denominator (2,800 sites, 2023). We print both side by side and refuse to average them, because averaging two different fractions produces a third number that measures nothing.

What percentage of sessions view a product page?

Nobody has published a credible benchmark for this, and we are not printing one. Google's GA4 documentation defines the view_item stage but publishes no benchmark figures at all (fetched July 30, 2026). Shopify Analytics has no session-to-product-view report. Every figure circulating for this stage that we traced led to a composite page with no sample size, no field date, and no primary source.

How do I find my own funnel drop-off in Shopify?

Open Analytics, then Reports, then the behaviour reports. Shopify publishes three session metrics you need: sessions with cart additions, sessions that reached checkout, and sessions that completed checkout (Shopify Help Center, fetched July 30, 2026). Divide each by the previous one for your three transition rates. Two traps: Shopify's session metrics only have historical data from October 1, 2022 onward, and reached checkout counts sessions with user input during checkout, not arrivals.

Why do Shopify and GA4 show different funnel numbers?

Because they count different things. Shopify counts sessions. GA4 counts users, in a closed funnel where users are only counted in steps they complete in the specified sequence. Shopify's reached checkout fires on user input during checkout, while GA4's begin_checkout is an event. Both are documented on the official help pages, and neither is wrong. They are simply not the same measurement, so never benchmark one against the other.

How much revenue is one bad funnel stage costing me?

Multiply your monthly sessions by the gap between your stage rate and the benchmark, then by your own downstream stage rates, then by average order value and contribution margin. Treat the result as a ceiling for prioritising your four stages, not a forecast. Visitors who need persuading to add to cart are by construction lower intent than those who added unprompted, so they will not convert downstream at your existing rates.

Sources

Every number in this post traces to one of the sources below. Pages were fetched, or tool calls made, on July 30, 2026. Rolling benchmarks change over time; figures are as read on that date.

1. DTC Pages, “Ecommerce Conversion Rate Benchmarks 2026” – session to add to cart 5.95%, 25% drop to checkout, 51.6% checkout abandonment, 2.16% mean conversion rate, device splits. 21 Shopify stores, $417M combined revenue, 179M+ sessions, data January 2025 to June 2026. Published April 4, 2026. Fetched July 30, 2026.
2. Littledata, checkout conversion rate benchmark – 45% average checkout completion, mobile 44%, desktop 49%, top 20% above 59%, top 10% above 66%. 2,800 ecommerce sites, 2023 study. Fetched July 30, 2026.
3. Littledata, add-to-cart rate benchmark – 4.6% average, top 20% above 7.5%, top 10% above 9.6%, fashion 5.4%, food and beverage 4.8%. Denominator not disclosed on the page. 2,800 sites, 2023 study. Fetched July 30, 2026.
4. Dynamic Yield, add-to-cart rate benchmark – 6.07% global, regional, device and vertical cuts. Defined on page as items added to cart after product page views. Rolling trailing 12 months; the value moves. Read July 30, 2026.
5. Dynamic Yield, cart abandonment rate benchmark – 77.54% global, regional and device splits, Beauty 79.81% high and Pet Care 51.1% low. Rolling 12 months. Read July 30, 2026.
6. Dynamic Yield, ecommerce benchmark hub – panel description (200M+ monthly unique users, 400+ brands, 300M+ sessions), device split, conversion rate and average order value by vertical. Read July 30, 2026.
7. Baymard Institute, cart abandonment rate statistics – 70.22% meta-average across 50 studies, the 42% “just browsing” figure, and the reason distribution excluding browsers. Page states last updated September 22, 2025. Fetched July 30, 2026.
8. Baymard Institute, checkout usability research – 70.19% cart abandonment on the same day, 335 sites benchmarked, 30,000+ elements reviewed, 65% mediocre or worse, 2% good, 32 average improvements, 35% modeled potential improvement. Fetched July 30, 2026.
9. Baymard Institute, “Checkout Optimization: From 16 Fields to 8” – 12.8 average checkout form fields, 6 to 8 achievable, 14% improvement on the 14.88 average measured in 2016. Based on 7,900+ ratings of the world's 60 largest ecommerce sites. Published June 21, 2019. Fetched July 30, 2026.
10. eMarketer, “Extra costs are the top reason consumers abandon online carts” – 48% of US adults, sourced to Baymard Institute, n = 1,012 US adults 18+, fielded February 7, 2024. Article published June 7, 2024. Fetched July 30, 2026.
11. Shopify Help Center, behaviour reports – verbatim definitions of sessions with cart additions, sessions that reached checkout, sessions that completed checkout, and conversion rate over time, plus the October 1, 2022 historical-data note. Fetched July 30, 2026.
12. Google Analytics Help, purchase journey report – the five purchase-journey stages, user-counted closed funnel, and the worked abandonment example. Publishes no benchmark figures. Fetched July 30, 2026.
13. Contentsquare, 2026 Digital Experience Benchmark – retail conversion rate down 5.1% year over year. 6,500+ websites, 99 billion sessions, Q4 2024 to Q4 2025. Fetched July 30, 2026.
14. Contentsquare, conversions benchmark guide – returning 2.9% vs new 1.7%, 52.8% return visits, mobile 69.9% of traffic, paid search 2.8% and organic social 0.7%. 99 billion sessions, 6,000+ sites, Q4 2024 to Q4 2025. Fetched July 30, 2026.
15. Contentsquare, 2025 Digital Experience Benchmarks press release – 53% single-page exits, +10% session depth associated with +5.4% conversions, paid-social traffic effects. 90 billion sessions, 6,000 websites. Dated January 28, 2025. Fetched July 30, 2026.
16. IRP Commerce, ecommerce market data – UK sector conversion rates for June 2026. Confirmed by fetch to publish no add-to-cart or cart abandonment rate. Fetched July 30, 2026.
17. Triple Whale, ecommerce benchmarks – $74.12 median AOV (its own panel, paid channels, 2025), the Statista attribution for 70.22%, and the Blend Commerce citation for “approximately 7.5%”. Page states last updated February 26, 2026. Fetched July 30, 2026.
18. Blend Commerce, add-to-cart rate – the intermediate hop in Chain A, reporting Littledata as 4.4% average and 7.6% as the top-20% threshold, with no date or link to the original. Fetched July 30, 2026.
19. Store Growers, “40+ Ecommerce Metrics Benchmarks (2026)” – cited only as evidence of staleness: page last updated 8 January 2026, serving Monetate Q2 2018 add-to-cart data and Baymard 2020's 69.80%. Fetched July 30, 2026.
20. Shopify blog, ecommerce funnel – cited for the absence of stage benchmarks and the one million / 500,000 hypothetical. Dated March 4, 2023. Fetched July 30, 2026.
21. Heap, ecommerce conversion funnel optimization guide – cited for the absence of stage benchmarks; carries a 2.86% overall conversion rate attributed to Invespcro (2020) and a 2019 IRP Commerce table, with no publication date displayed. Fetched July 30, 2026.
22. DataForSEO serp_organic_live_advanced, query “ecommerce conversion funnel”, United States, English, desktop. The page-one composition and result dates quoted in this post. Called July 30, 2026.
23. DataForSEO ai_optimization_chat_gpt_scraper, United States, English. The ChatGPT funnel-benchmark answer, its three citations, and the quoted framing. Called July 30, 2026.
24. DataForSEO backlinks_summary and backlinks_bulk_ranks. Referring domains, first-seen dates, crawled page counts, spam scores and domain ranks for the cited domains and for baymard.com, contentsquare.com, littledata.io, marketing.dynamicyield.com and dtcpages.com. Called July 30, 2026.
25. DataForSEO dataforseo_labs_google_keyword_overview, United States, English. Search volume, keyword difficulty and average referring domains for the terms discussed. Called July 30, 2026.