Guides

What 2,603 reviews say actually goes wrong with inventory sync

Original research: every public App Store review of the nine main Shopify sync apps. Not one merchant in 156 complained about what they all market on.

Updated 28 August 2026

Every app in this category sells the same promise: real-time sync. So we read every public review of all nine of them — 2,603 reviews — to find out what merchants say when it goes wrong.

Not one of them complained that the sync was too slow.

That is the finding worth sitting with. Of 156 low-star reviews containing written text, zero mentioned lag, delay, or waiting for a sync. The category competes on the one dimension its unhappy users never mention.

This is our own research and we are one of the apps in the market, though not one of the nine studied — StockUnison launched after the crawl and has no reviews to include. The method is below so you can disagree with it.

What we did

On 16 July 2026 we crawled the public App Store review pages for nine apps: syncX: Stock Sync, QuickSync All-in-One, Trunk, Syncio, Multi-Store Sync Power, Tipo Multistore Sync, Sumtracker, SyncLogic and Horse Inventory. That is 2,603 reviews, every one publicly visible on those listings at that moment.

A third of that corpus is no longer public, and it matters. Re-reading the same nine listings on 28 August 2026 gives 2,272 reviews — 331 fewer, about 13%, removed by Shopify’s mid-2026 review-policy enforcement. It did not fall evenly: SyncLogic lost 45% of its reviews, Multi-Store Sync Power 28%, Syncio 24%, while Sumtracker and Synkro lost none at all. So the per-app counts below describe July, and a listing you open today will show fewer. The findings themselves do not move — a complaint someone wrote is still a complaint someone wrote — but two things are worth saying. Where an app’s complaint share rose across the sweep (Multi-Store Sync Power 11.3% to 13.4%, SyncLogic 16.8% to 19.2%), it lost proportionally more positive reviews than negative ones. And a review count is a weaker signal than it looks when part of any total is simply what survived a policy sweep.

Ratings, dates, countries and the merchant’s stated usage duration come from the listing markup. The complaint classification is ours: keyword patterns applied to the text of reviews rated three stars or below. Where a review mentions two problems it is counted in both themes.

The shape of the data

Merchants broadly like these tools. 91% of all 2,603 reviews are five stars, and only 163 (6.3%) are three stars or below. 138 reviews carry a rating with no text, leaving 156 low-star reviews whose complaints can be read.

Average rating separates nothing: every app in the set sits between 4.46 and 4.92. The complaint share separates them enormously — from 1.1% to 16.8% across the same nine apps. If you are comparing apps on the star average, you are reading the one number that carries no information.

What actually goes wrong

Share of the 156 low-star reviews with text mentioning each theme:

Three things stand out.

Money is the most common complaint, not stock. Unexpected charges, price increases and refund disputes outrank the operational failure the entire category is sold on. This is worth knowing before you choose on pricing axis — per store, per product or per variant — because the axis is what produces the surprise.

Data loss is as common as overselling. What actually happens when you disconnect is the guide built from these reports. Deleted products, wiped catalogues and reset quantities appear about as often as the failure everyone anticipates. Merchants expect a sync app to sometimes get a number wrong; they do not expect it to remove things. Ask any vendor what disconnecting changes, in writing, before you install.

Nobody complains about speed. Every listing in this category leads with real-time. Zero unhappy merchants raise it. The plausible reading is not that speed is unimportant but that it is table stakes — everyone is fast enough, so being faster wins nothing, while being wrong or expensive loses customers. What “real-time” actually has to mean is a mechanical question about conditional writes and reconciliation, not a stopwatch one.

The finding that should change how you evaluate

These tools fail late.

Merchants who had used the app for over a year are roughly 1.6× over-represented among the complaints. Every other tenure bracket is under-represented.

That inverts the usual assumption that trouble shows up during onboarding. It fits what this category’s failure mode actually is: a sync that silently stops matching after a catalogue edit, a webhook subscription that quietly lapses, a duplicate SKU introduced during a seasonal relist. None of that appears in week one. It appears in month fourteen, after you have stopped checking — which is the argument for a scheduled drift check rather than a one-time setup verification.

It also means “it’s been fine for six months” is weak evidence. The reviews say the risk is ahead of you, not behind.

Complaint rate is rising

Roughly double the 2018–2021 baseline. The data shows the trend and cannot explain it — declining quality, rising expectations, larger catalogues and more complex multi-store setups are all consistent with it.

By app

Review counts are mostly age and market size, not quality — Stock Sync leads because it launched in 2014 and serves the much larger supplier-feed market. Read the complaint share instead, and note that small bases move fast: Horse’s 35 reviews carry far less evidence than Trunk’s 443.

Two apps in this category are missing from the table on purpose. Synkro (89 reviews, not one of them at three stars or below) and Syncerize (13, likewise) have no unhappy reviews to read, which is the thing this study measures. That is not a clean bill of health and it is not a black mark: on a base that size, a complaint share of zero and a complaint share of five percent are the same evidence. Read their listings for what the app does; do not read the absence of complaints as proof there are none coming.

What to do with this

1. Compare on complaint share, not star average. The averages are indistinguishable; the complaint shares differ by 15 percentage points.
2. Ask about billing before features. The most common complaint in the category is financial. Establish the pricing axis and what happens at a tier boundary.
3. Ask what disconnecting changes, in writing. Data loss is as common as overselling and far harder to undo.
4. Discount speed claims. Zero unhappy merchants raised it. Ask instead whether writes are conditional and what reconciles a missed webhook.
5. Plan for late failure. Run a drift check on a schedule, not once at setup. The reviews say year two is when it bites.

Limits of this study

The classification is keyword-based, so a complaint phrased unusually is missed and a review mentioning a word in passing may be over-counted; the themes should be read as indicative rather than exact. Only reviews still publicly visible on 16 July 2026 are included — removed or filtered reviews are invisible to any crawl. That limit turned out to be a large one: 13% of this corpus has been unpublished since, as set out above, which is a reminder that any review study measures what a platform was willing to show on one day. Reviews are self-selected, skewing toward the delighted and the furious. And the nine apps are not the whole market.

Numbers move. The listings are linked throughout so you can check today’s.

Why we published it

Partly because it is useful and nobody else had done it. Partly because it is the argument for how StockUnison is built: nothing is ever deleted, pricing is flat per connected store rather than per product or variant, and the difference report exists specifically because the reviews say failures arrive in year two rather than week one.

We are not in the table. StockUnison launched in August 2026 and has no reviews to include, which is a disadvantage this research does nothing to fix.