Back to E-commerce Tips

Post-Purchase Upsell Metrics: What to Track Beyond Take Rate

E-commerce Tips
Post-Purchase Upsell Metrics: What to Track Beyond Take Rate

Once your first post-purchase offer is live, the obvious question is whether it works. The obvious answer, "what share of buyers said yes", is only one of several post-purchase upsell metrics, and on its own it can point you the wrong way. A high take rate on a tiny add-on can earn less than a modest take rate on a better product, and a great take rate means little if most buyers never see the page.

This guide covers the numbers worth tracking, how to calculate each one, and what each tells you to change.


Start with reach: who actually sees the offer

Before you judge an offer, check how many orders it reached. Post-purchase offers lose buyers at several points before anyone clicks:

  1. Eligible checkouts. Shopify only shows a post-purchase page for some checkouts. Wallets like Apple Pay and Google Pay, pay-later services, gift cards and checkouts in another currency skip it (Shopify's post-purchase docs).
  2. Matched an offer. Of the eligible checkouts, how many matched your targeting? An offer that only targets one collection will skip everything else.
  3. Shown. Of the matched checkouts, how many actually displayed the offer? Sold-out products and slow page loads cost you a few here.

Each step as a share of the one before tells you where to look. A big drop at the first step is about how your buyers pay, not about your offer, so no amount of copywriting fixes it. A big drop at the second means your targeting is narrower than you think. If you're losing most orders before the page, see why a post-purchase upsell isn't showing.


Take rate: the headline number, read correctly

Take rate (also called accept rate or upsell conversion rate) is:

accepted offers ÷ offers shown

The denominator matters more than people expect. Some articles divide by all orders, some by offers shown, some by checkouts that reached the page. The same store can report three different "take rates" from the same week. That's also why benchmark numbers you find online rarely compare cleanly with yours: you don't know what they divided by.

Pick one definition, write it down, and compare only against your own past numbers. "Accepted out of shown" is the most useful for judging the offer itself, because it ignores orders the offer never had a chance with.

A few things move take rate without making the offer better:

  • A lower price. Cheap add-ons get more yeses but may add less revenue.
  • A bigger discount. More acceptances, smaller margin on each one.
  • Narrower targeting. Showing the offer only to the best-matched buyers raises the rate and lowers the reach.

None of these are bad, but they're trade-offs, which is why take rate needs the next metric beside it.


Revenue per shown checkout: the number that settles arguments

If you only track one number besides reach, make it revenue per shown checkout:

upsell revenue ÷ offers shown

It combines take rate and offer value into one figure. An offer accepted by 1 in 10 buyers at $8 earns $0.80 per shown checkout; one accepted by 1 in 20 at $25 earns $1.25. The second has half the take rate and earns more.

This is the right number for comparing two versions of an offer, two products or two discount levels, because it rewards whatever actually brings in money. When you A/B test post-purchase offers, compare the versions on this, not on take rate alone.

Be clear about what "revenue" includes. Gross revenue, with tax and the shipping the offer added, is what the buyer paid. Net revenue, after discount and before tax and shipping, is closer to what you keep. Pick one and use it consistently.


What the discount costs you

A discount is the most common reason buyers say yes after checkout, and the easiest way to give away margin. Track the average discount per accepted offer next to take rate. When you raise a discount, take rate should rise enough to pay for it; check revenue per shown checkout (in net terms) before and after.

If a deeper discount only nudges take rate, the buyers weren't declining because of price. Try a different product or better targeting instead. For setting the discount in the first place, see how much discount to give on a post-purchase upsell.


Declines, timeouts and failures

Not every "no" is the same. Split the outcomes:

  • Declined: the buyer saw the offer and clicked no. This is normal and most buyers will do it.
  • Timed out: the buyer left the page or let the time run out without answering. A high share can mean the page is too long or loads slowly on mobile.
  • Failed: the buyer accepted, but the payment or order update didn't go through. Usually rare, but worth watching, because each one is a buyer who said yes and didn't get it.

If you use a downsell after a decline, track it separately. A downsell that converts well tells you the first offer's price or product was close but not quite right.


Metrics that live outside your upsell report

Some of the most important post-purchase upsell metrics never show up in the offer report:

  • Returns and refunds on upsold items. If buyers send the add-on back, the offer is a bad match, however good its take rate. Tag or filter upsold orders in Shopify and check their return rate monthly.
  • Support tickets. Buyers who didn't realize they'd been charged twice, or who wanted a different variant, will tell you. A spike after a change is a signal to undo it.
  • Average order value. Upsells raise it, but so do promotions and seasonality. Use AOV as a sanity check, not as proof that an offer works.

One trap to know about: Shopify notes that third-party analytics services report only the initial purchase's event and value (Shopify's post-purchase docs), not what the buyer added afterwards. Don't judge your upsells in your ad or analytics dashboards. Measure them in your upsell app's own reports, and check your order totals in Shopify if the two seem to disagree.


A monthly review of your post-purchase upsell metrics

You don't need a dashboard habit. Once a month:

  1. Check reach: eligible, matched, shown. Did any step drop?
  2. Rank your offers by revenue per shown checkout. Keep the top ones, rework the bottom one.
  3. Look at average discount next to take rate for any offer you changed.
  4. Check returns on upsold orders.
  5. Change one thing, and give it enough orders before you judge it.

Changing one thing at a time is what makes the numbers mean anything. Change the product, discount and targeting in the same week and you won't know which one did it.


Where Elite Post Purchase fits

Elite Post Purchase's Analytics page is built around these numbers. Its Overview follows each checkout through Eligible checkouts, Matched a flow, Offers shown and Offers accepted, with each step's share of the one before. Beside it are Gross revenue and Net revenue, and a Detailed breakdown per flow, offer and product with Accepted, Declined, Timed out, Failed, Accept rate and Avg. discount. A Why nothing showed section flags fixable reasons a flow was skipped. Read your analytics walks through every number, and A/B test a flow compares two versions on gross revenue per checkout.

Orders with an accepted offer are tagged ElitePostPurchase in Shopify, so you can filter them for the returns check above. See the features page for the rest.

Measure first, then change one thing. Your own numbers, read the same way each month, beat any benchmark.

E-commerceAnalyticsPost-PurchaseUpsellingConversion