Why apps accumulate without anyone deciding

Installing an app takes a few minutes. It needs no developer, no budget discussion and no sign-off. That is exactly what makes the App Store so useful, and exactly what causes the problem. Every installation is a small decision that nobody perceives as a decision about the shop’s architecture.

A typical pattern unfolds over two or three years. At launch, reviews and a newsletter form are added. The first discount campaign brings a countdown timer. Later come an upselling widget, a size advisor and a product filter app. Some of these stop being used once the campaign ends, yet they stay installed.

Anyone opening the app overview today often finds entries whose purpose nobody can quite name. That would be harmless if an app only cost its monthly fee. It costs in four places, however, and only one of them appears on the invoice.

The four costs of an app, only one of which is on the invoice

Load time. Many apps load their own JavaScript, stylesheets and fonts into the shop, often from third-party servers. The browser has to request and process each of these files before the page is fully usable. A single script barely registers. Five or six scripts loading on every page, including pages where their feature never appears, noticeably slow down mobile devices in particular.

Running costs. The fee is rarely fixed. Many apps charge by orders, revenue or contacts. What was cheap at launch grows with the shop, without anyone revisiting the decision. Billing usually runs through the combined Shopify invoice, where individual items are easily overlooked.

Data dependencies. Reviews, loyalty points, subscriptions or wishlists often live not in Shopify but in the app provider’s database. The shop then owns this data only as far as the app offers an export. Data protection adds another layer. If an app processes customers’ personal data, its provider is a processor under the GDPR and needs a corresponding agreement.

Maintenance. Apps update themselves; the theme does not keep the same pace. An app update, a theme change or a change on Shopify’s side can cause display errors. Two apps that modify the same element, such as the cart, also tend to conflict. Troubleshooting then takes time, and providers are quick to point at each other.

Reducing Shopify apps starts with measuring each app

An overall score from PageSpeed Insights shows that a shop is slow. It does not show why. Deciding about individual apps requires the contribution of each one. Why load time matters commercially in the first place is covered in the article on website performance.

A practical route runs through a copy of the theme. In it, app integrations, called “app embeds” in Shopify, can be switched off one by one in the theme editor without touching the live shop. The same three pages are then measured again: the home page, a product page and the cart. Because measurements in testing tools fluctuate, several runs per state are worthwhile. The Core Web Vitals, Google’s metrics for loading and responsiveness, provide orientation, such as 2.5 seconds until the largest visible element has loaded.

The network tab in the browser’s developer tools helps as well. Sorted by domain, it shows which providers load files on a page. If a provider appears whose app was uninstalled long ago, that leads straight to the problem in the next section.

Alongside this, a simple list takes shape. It records every app with its purpose, monthly cost, billing model, responsible person and the place where its data is stored. Any app for which neither a purpose nor a responsible person can be entered is already a candidate for removal.

What stays behind in the theme after uninstalling

The obvious assumption is that an uninstalled app has gone. For newer apps, that is largely true. They hook in through theme app extensions, building blocks that Shopify manages and removes together with the app.

Older apps and some installation guides work differently. They insert code snippets directly into theme files, for instance into the main layout file. These snippets remain after the app is uninstalled. At best, they load nothing and cost a single request. At worst, they cause errors in the browser or empty areas on the page.

An app’s metafields, custom pages and discount rules sometimes outlast its removal as well. A search of the theme files for the provider’s name reveals such leftovers. A backup of the theme belongs before any deletion. And before uninstalling an app that holds its own data, the export always comes first, because afterwards it is often no longer possible.

App, custom code or nothing: the criteria for deciding

Cleaning up only helps if future features are decided differently. Every desired feature has three routes: an app, custom code in the theme, or doing without. A fourth comes before them and is often overlooked.

Start with what is built in. Metafields and metaobjects, freely definable extra fields for products and content, replace many apps for size charts, ingredient lists or product notices. Shopify Flow automates simple workflows, and filters can be set up with Shopify’s own search and filter app. Whatever is covered here needs no third-party provider.

An app when the feature is standard and complex. Subscriptions, payment methods or tax logic are hard to build correctly and change with regulation. A specialised provider carries that risk better than custom code. What matters then is whether the app keeps data exportable and integrates through theme app extensions.

Custom code when the feature is small and rarely changes. An announcement bar, a delivery-time note, a product badge: features like these take a few lines of Liquid, Shopify’s templating language. They load only where they are needed and carry no monthly fee. The price is that someone has to understand the code and carry it over when the theme changes.

Nothing when nobody uses the feature. Pop-ups, countdown timers or recently viewed widgets can be checked for usage in web analytics. If there is barely any interaction over several weeks, doing without is the cheapest and fastest solution. This is precisely what is hard, because the feature was once introduced on purpose.

Reducing Shopify apps: when the effort is not worth it

The obvious intuition is to keep free apps, replace expensive ones, and assume custom code is always lighter. Neither holds. Some free apps display their provider’s logo in the shop or load far more code than necessary. A paid app that integrates cleanly through theme app extensions may weigh the shop down less.

Custom code moves maintenance in-house. For a shop without regular access to a developer, that is a risk. If the code breaks after a theme update, no support ticket will help. An app with dependable support is then the more stable choice, even if it charges fees.

Switching has its own price too. Replacing an app with an established data set, such as a loyalty programme with point balances, means migrating data and informing customers. If the existing app loads quickly, charges fairly and releases its data, it is not a problem. Swapping it out merely to lower the number of apps is not worthwhile. The aim is not a short list, but one in which every entry is justified.

A lean shop only stays lean if someone maintains the list

After the clean-up comes the phase in which most shops start to grow again. New campaigns bring new apps, and the inventory list goes out of date. A fixed rule helps: every new app gets an entry with purpose, cost, data location and responsible person before it is installed. Once a quarter, the list is compared with the actual installed apps and a fresh measurement.

That keeps the app stack a deliberate decision rather than a sediment. How we look after existing shops and systems over time is described under AI Operations Care.