Five of seventeen apps
stopped telling you
what they cost.

What changed in Shopify app
pricing between July and
September 2026?

Twelve of seventeen tracked apps changed something about their published pricing in roughly sixty days. I checked each one against the vendor's own pricing page and its Shopify App Store listing on 20 September 2026, which is the same method behind the Shopify app pricing benchmarks. Five published less than before, and five published more.

That is a lot of movement for a category that used to reprice about once a year. It is also lopsided in a way that is easy to misread. If you only look at the five that went quiet, the story is that apps are going upmarket. If you only look at the five that got cheaper or clearer, the story is a price war. Both readings are half the data.

The two I could not verify are worth naming rather than rounding off. Klaviyo rebuilt its pricing page as a plan builder during the period, so only the free tier survives as static text. Yotpo's loyalty listing did not resolve at the slugs I tried, and I could not establish whether it moved or was withdrawn. Neither absence is evidence of a price change, and neither is evidence against one.

Which apps stopped publishing
prices, and what replaced
the numbers?

Five apps removed published pricing. Triple Whale went furthest, replacing named-price tiers with four packages that all route to sales. The other four kept a published entry price and withdrew the tier above it, which is the quieter and more common version of the same move.

Skio's is the one to watch, because it has a cause you can name. Recharge acquired Skio in April 2026, and by September the $399 tier that sat below Recharge's own $499 plan had gone. A cheap entry point that undercuts the parent company rarely survives the integration. The published floor for a small subscription brand rose by $200 a month without a price increase on any surviving plan.

Okendo's version is subtler and probably more instructive. It still publishes $19, $119 and $299, so nothing looks withdrawn at a glance. What changed is the ceiling: the tier that used to carry a number at 10,000 orders is now sold as Okendo Platform and Platform Scale, both quoted. You can still price the bottom of the ladder. You can no longer price the top.

Why did the other half publish
more instead of less?

Because they are selling a different kind of thing. Smile.io cut its entry tier from $49 to $15 and inserted a $79 step beneath its $199 plan, so the bottom of the loyalty category is now cheaper than it was in July. Smile is buying back the bottom of the market.

ReConvert did something more interesting. Its old card ran $4.99, $7.99 and $14.99 by order band, with 0.75% of app-generated revenue taken on the top tier and a separate revenue-based plan from $50. The new card is four plain order bands, free up to 50 orders, then $9.99 to 2,000, $199 to 10,000 and $499 to 20,000, with no published revenue share at any tier. The prices went up. The disclosure got simpler.

Yotpo Reviews moved the other way from where you would expect too, replacing "custom pricing at volume" with published Starter and Pro tiers. Postscript now prints its per-message rate at every tier rather than a range. Loox finally wrote down the part of its Convert plan that used to surprise people: $49.99 covers 300 orders, and each additional 300 costs $50.

Worth flagging one contradiction inside that group, because it changes how you negotiate. Yotpo publishes two different prices for the same plan names. Its Shopify listing shows Starter at $15 and Pro at $119. Its own pricing page shows $89 and $169 at the default order volume. Both are Yotpo's numbers, on Yotpo's surfaces, on the same day.

The dividing line is the
word platform.

Three of the five apps that went quiet now use the word by name. Okendo sells Okendo Platform and Platform Scale. Rebuy quotes a Platform One rate. Triple Whale sells packages rather than plans. The pattern underneath is cleaner than the billing model, and it predicts the behaviour better than company size does.

Here is the rule as I would state it. An app keeps publishing its price while the top tier is the same product with a bigger number attached. It stops publishing when the top tier becomes a different product. The moment a vendor bundles, the price stops being a lookup and starts being a configuration, and a configuration has to be quoted.

Test it against the case that should break it. Okendo and ReConvert are both order-tiered, both mid-market, both moved in September. One went quiet above 3,500 orders. The other publishes all the way to 20,000. The difference is not the billing model, which is identical. Okendo's top tier bundles reviews, loyalty, quizzes, referrals and surveys into one platform. ReConvert's top tier is the same upsell app with a larger cap.

That reframes what the withdrawals are telling you. What they announce is a product decision: the vendor has decided its product is a suite, and suites cannot be listed on a card. If you want to predict which app disappears from the published benchmark next, do not look at who raised prices. Look at who shipped a second product.

Is the ecosystem moving
upmarket, or is sales
moving down?

Both, and the second half is the part that gets missed. The price ceiling did go up. But the threshold at which you stop being allowed to self-serve came down at the same time, and for a mid-sized brand that is the change that actually bites.

Run the numbers on where the sales conversation now starts. Okendo routes to a quote above 3,500 monthly orders. At an $80 average order value, a figure I am supplying rather than the vendor, that is roughly $3.4M a year. LoyaltyLion routes to a quote above 500 monthly orders, which on the same assumption is around $480,000 a year.

Neither of those is an enterprise buyer. A brand doing under half a million a year in revenue is now meeting a salesperson to buy a loyalty program. Sales coverage is expanding downward, into a segment a pricing page used to serve on its own, and it only works because something else is doing the qualifying.

The free tiers are doing it. Thirteen of the seventeen tracked apps run one, and four of the five that went quiet have a free plan sitting under the quoted tiers. The free tier earns its place by qualifying. It gets the integration installed, and the integration tells the vendor your order volume and your GMV before you have seen a number from them.

What breaks for a merchant who
can no longer look up
a price?

Four things break, and only the first one is obvious. Across the brands I have operated and advised, a working app stack runs 1% to 3% of revenue once the order-tiered, contact-based and revenue-share tools are counted together. You cannot forecast that band from public pages any more, and the failure compounds.

1. Year two stops being forecastable. You can price the tier you are on. You cannot price the tier your growth plan puts you in, which is the number that actually belongs in the model. A brand planning to triple orders now has an unpriced line in its operating budget.
2. The vendor sees your numbers before you see theirs. A free tier installs an app that reads your order volume and GMV. By the time you ask for a quote, the other side has better information about your business than you have about their pricing. That asymmetry did not exist when the price was on a page.
3. Renewal leverage inverts. A published ladder caps what a renewal can ask for, because you can see the next rung. A quoted plan has no visible ceiling and usually sits on an annual contract, so the uplift conversation starts from whatever the vendor proposes.
4. Comparison stops working. You cannot run a shortlist against a vendor that publishes nothing without booking three demos first. The practical effect is that quoted vendors get evaluated less often, which is not an accident.

The countermeasure is unglamorous and it works. Before you sign anything quoted, get three things in writing: the full tier ladder with the thresholds that move you between rungs, the overage rate once you pass your cap, and a cap on the renewal uplift. If a vendor will not put the ladder in writing, you are not buying software, you are buying a negotiation every year. That is the same exposure as any other platform dependency you cannot price. The same discipline applies to the wider cost of app bloat in a growing stack.

What should you do if you are
setting your own app's
price?

Treat publishing as a positioning decision rather than an operational one, because that is what buyers now read it as. With five of seventeen category leaders going quiet in a quarter, a published price has stopped being the default and started being a statement about who you sell to.

1. Publishing says self-serve, hiding says sales team. Neither is wrong, but you should pick it deliberately and staff for it. A hidden price with no sales team is the worst of both: you have removed the self-serve path and not replaced it.
2. Hiding your price removes you from every comparison an AI assistant can build. This is the cost nobody prices in. Ask an assistant what attribution costs in 2026 and it can quote Judge.me at $15 because Judge.me publishes it. It cannot quote Triple Whale, because there is no number to extract. A quoted vendor is absent from the answer, not ranked lower in it.
3. Size the free tier at the point where you want the conversation. The free tier is the qualifier now, not the trial. LoyaltyLion's sits at 400 monthly orders and Loox's at 500, and those numbers are a statement about where each company thinks a human should get involved.
4. If you go quiet, go quiet above a threshold. Okendo kept $19, $119 and $299 published and withdrew only the top. It keeps the self-serve funnel, keeps the citable numbers, and still routes the bundle to sales. Withdrawing everything, as Triple Whale did, gives up the whole top of the funnel for the same commercial benefit.

The second point is the one I would think hardest about. It cuts against the instinct that hiding a price protects margin, and on a two-year view it may be the more expensive choice. How a category gets summarised by assistants is becoming a distribution channel, and an app with no published number cannot appear in the summary. The model you choose for your own app now carries a visibility cost as well as a revenue one.

Where this leaves the benchmark,
and what I would watch
next.

The benchmark is now two datasets in one. It records what apps charge, and it records which apps will still tell you. The second one has become the more interesting column, because it moved faster in sixty days than most of the prices did.

Gorgias is a useful counter-example. It still publishes every tier from $10 to $900, and its complications sit in the AI and overage lines, well away from the plan card. Three things are worth watching on the next pass. Klaviyo is the largest line in most DTC stacks and it has just moved its pricing behind a plan builder, so whether the profile tiers survive as published numbers matters more than any single change above. Recharge now owns Skio and has already removed one tier. And nobody has reversed course yet, which is the thing that would tell you this is a cycle rather than a direction.

For what it is worth, I do not think this is a pricing story at heart. It is a product story showing up on a pricing page. Apps that shipped a second product had to stop publishing, because a bundle has no single number. It is the same consolidation pressure mapped in the 2026 Shopify app category map. The cost of it lands on the brands in the middle, who are too large to self-serve and too small to negotiate well.

If you are in that middle, the practical move is to build your own benchmark before you need it. That habit is most of what I do on a standing advisory engagement. Keep a dated record of what every vendor charges, and the renewal conversation starts from evidence instead of from whatever the vendor proposes. Record what each vendor in your stack published on the day you signed, and what it publishes now. That record is your only leverage in a renewal against a vendor who has stopped publishing anything, and it takes about an hour a quarter to keep.

Questions founders and
merchants ask when a
price disappears.