E-commerce App Marketing: How to Acquire Shoppers Who Actually Buy

E-commerce app marketing is the practice of acquiring, activating and retaining shoppers for a mobile commerce app, with the first order and repeat purchases as the goal rather than the install. It covers paid user acquisition, creative, app store conversion, deep linking and lifecycle work for shopping, marketplace, deals and delivery apps. An e-commerce app is any mobile app whose revenue comes from transactions inside the app: retail, groceries, food delivery, classifieds, deals or gifting.

The difference from other app categories is simple. A subscription app earns its money from one decision: start a trial, convert to paid. A shopping app earns its money from many small decisions spread over months: browse, add to cart, check out, come back next week. That changes which event your campaigns should bid on, which creative wins, how the store page should look and how you measure success. This guide sets out how the Admiral Media team approaches e-commerce app marketing, based on our work managing over €500M in mobile ad spend for 150+ mobile brands, with results from shopping, delivery, marketplace and gifting apps we have scaled.

Why e-commerce apps need a different user acquisition model

E-commerce apps need a different user acquisition model because the install is the cheapest and least predictive event in the funnel. A shopper who installs during a sale and never orders costs the same to acquire as one who orders every week, so any model that optimizes toward installs alone will fill the app with low-value users.

Three structural traits make commerce apps behave differently from games or subscription apps.

Revenue arrives in many small transactions

Lifetime value in a commerce app is built from order frequency multiplied by basket size multiplied by margin. None of those three is visible on day zero. A subscription app can see a trial start within minutes of install. A grocery or fashion app may not see a first order for days, and the second order is the real signal of a retained customer. That gap is why Admiral Media treats the first order as the minimum viable optimization event and the second order as the true north metric for mobile app LTV modeling.

GMV is not revenue, and revenue is not margin

Commerce apps report gross merchandise value, but the business keeps only a commission, a markup or a delivery fee. Two campaigns with identical ROAS on GMV can have completely different economics if one attracts shoppers who buy discounted, low-margin items and the other attracts full-price buyers. Bidding on GMV without adjusting for margin is one of the most common reasons a commerce app scales spend while losing money on every cohort.

Promotions distort every signal

First-order discounts, free delivery and seasonal sales pull forward demand and inflate early conversion rates. They also attract deal seekers who churn as soon as the discount ends. In Admiral Media’s experience, the cohorts acquired with the heaviest launch discount are rarely the cohorts with the best 30-day repeat rate. Promotions are a legitimate lever, but they need to be measured on the cohort after the promotion ends, not during it.

These traits explain why e-commerce app marketing is closer to retail economics than to app install marketing. The rest of this guide is built around that idea.

The Admiral Media First-Order Funnel Framework

The Admiral Media First-Order Funnel Framework is a six-step method for scaling a commerce app by moving campaign optimization deeper into the funnel as data volume allows, from installs to first orders to repeat orders. It is the structure the Admiral Media team uses when taking ownership of user acquisition for shopping, delivery and marketplace apps.

The First-Order Funnel Framework by Admiral Media

1. Define the unit economics first. Before launching a campaign, agree on contribution margin per order, average orders per customer in the first 30 days, and the maximum cost per first order the business can afford. This number, not CPI, becomes the north star for every channel.
2. Instrument the funnel end to end. Track install, registration, first add-to-cart, first order and second order as distinct in-app events, passed to your attribution partner and to each ad platform. Without separate first-order and repeat-order events, bidding algorithms cannot tell a browser from a buyer.
3. Start shallow, then move deeper. Launch with an event the campaign can hit often enough to learn, typically installs or registrations, and move to first-order optimization as soon as conversion volume supports it. This is the sequence Admiral Media used with a UK quick-commerce grocery client: installs first, then sign-ups, then first orders.
4. Match creative to shopping intent. Build creative around concrete products, prices and use cases rather than abstract brand claims. Creator-led, product-in-hand content and dynamic product creative are the two formats that consistently earn their budget in commerce accounts.
5. Convert the store visit with intent-specific pages. Route each audience and keyword theme to a custom product page that shows exactly what that shopper came for, and deep link the install to the matching in-app screen.
6. Close the loop on the second order. Use owned channels, deep-linked re-engagement and catalog retargeting to turn first-time buyers into repeat customers, and judge acquisition channels on repeat-order cohorts, not first-order volume.

The framework is sequential by design. Each step depends on the one before it: you cannot optimize toward first orders until the event is instrumented, and you cannot judge channels on repeat-order cohorts until you know your unit economics.

Choosing the right optimization event for a shopping app

The right optimization event for a shopping app is the deepest event that still fires often enough for the ad platform to learn from it. For most commerce apps that means starting on installs or registrations and moving to first purchase once conversion volume allows.

Ad platforms learn from conversion volume. Bid toward an event that is too rare, and the campaign never exits the learning phase, spends erratically and delivers volatile costs. Bid toward an event that is too shallow, and the algorithm finds the cheapest people to install, who are rarely the people who order. The art is moving deeper at the right moment. Admiral Media covers the mechanics in detail in our guide to app event optimization; the table below summarizes how each event behaves in a commerce context.

The add-to-cart row deserves a caution. Cart abandonment is the normal state of online shopping, not the exception: the Baymard Institute’s analysis of 50 studies puts the average documented online cart abandonment rate at 70.22%. That figure covers e-commerce broadly rather than apps specifically, but the implication holds for app campaigns: an add-to-cart is a weak promise of a purchase, so treat it as a temporary proxy and move to first purchase as soon as you can.

Case study: moving a quick-commerce app from installs to first orders

Admiral Media took ownership of user acquisition for a UK-based quick-commerce grocery delivery company operating in London, Amsterdam and Paris, built a creative production framework with twice-weekly ideation and accelerated testing, and moved optimization from installs to sign-ups to first orders as data accumulated. The account grew from zero to seven figures in monthly user acquisition spend. The published paid results were:

• -91% paid CAC: the cost to acquire a paying customer fell sharply as optimization moved down the funnel.
• +1,210% paid purchases: purchase volume grew far faster than install volume.
• +5,152% paid users and +2,178% paid conversions: growth concentrated on the events that matter to revenue.
• +64% paid installs: installs grew the least of any metric, which is exactly what you want when campaigns stop chasing cheap installs.

The pattern in those numbers is the point. Installs grew by 64% while purchases grew by 1,210%. A team optimizing toward installs would have reported a modest win. A team optimizing toward orders built a business. The full story is in the fast commerce case study.

Creative strategy for e-commerce apps

Creative for e-commerce apps works best when it shows real products, real prices and real people using the app, because shoppers decide on concrete value rather than brand promises. Creator-led video and dynamic product creative are the two formats that most reliably earn budget in commerce accounts.

Commerce creative has an advantage most app categories lack: the product is visible. A meditation app has to dramatize a feeling. A shopping app can show the exact jacket, the exact discount, the exact grocery basket arriving at the door. Creative that wastes this advantage on generic lifestyle footage usually underperforms.

Creator content beats polished brand creative for deals and discovery

Admiral Media ran TikTok user acquisition for kaufDA, the German deals app that helps shoppers find discounts on groceries, cosmetics, furniture and electronics. The team started with in-house creatives, then moved to influencer-produced content after initial tests, and tested messaging that ranged from tangible benefits such as price to intangible ones such as trust and simplicity. Within one month, creator ads compared with regular ads delivered:

• +731% overall user growth: creator content scaled the user base far beyond the regular creative.
• +146% user activity growth: the new users were active, not just installed.
• -18% CPI: creator ads were cheaper per install than the regular creatives.
• Almost 70 million impressions in Germany within one month.

The two creative approaches that won were inspiring, authentic, good-quality content aimed at shopping enthusiasts, and content built around the persona of the typical kaufDA user. The details are in the kaufDA case study. For a wider view of the format, see Admiral Media’s guide to UGC ads for mobile apps.

Dynamic product creative solves the volume problem

A commerce app with thousands of products cannot brief a separate video for every category. Dynamic creative optimization (DCO) assembles ads from a product feed and a library of creative elements, so the app can show relevant products at a scale no manual production team can match. It also delays creative fatigue, because the audience keeps seeing new combinations.

Admiral Media set up programmatic retargeting with dynamic banner creative for a leading South African food and grocery delivery app. The team connected a product feed, automated ad variation generation by product category and kept offers current with market conditions. The published results:

• -32% CPA: dynamic creative cost less per acquisition than static ads over the reporting period from August 14 to October 31.
• 2,140+ unique variations: generated from a feed of 160 different elements.
• 3x creative volume: compared with the static creatives delivered across the account’s history.
• +77% spend: budget scaled as performance held.

What makes a commerce ad convert

Across the commerce accounts the Admiral Media team manages, the creative patterns that tend to win share four traits:

• Product in the first seconds: the hook shows a specific product or basket, not a logo. Admiral Media’s guide to video ad hooks covers the opening seconds in depth.
• Price or value made explicit: a visible saving, a delivery time or a price comparison gives the viewer a reason to install now.
• The app experience on screen: real in-app footage of browsing, adding to cart and checking out reduces the perceived effort of switching from a website or a competitor app.
• One audience per creative: a deals hunter, a weekly grocery shopper and a gift buyer respond to different messages, so each concept should speak to one of them.

Converting the store visit with custom product pages and deep links

Custom product pages convert more shoppers because they show each audience the products and use cases it searched for, instead of one generic store listing. For commerce apps with distinct audiences, intent-matched pages are one of the cheapest ways to lower acquisition costs.

Apple allows an app to publish up to 70 custom product pages, each with its own screenshots, promotional text and app previews. Each page can carry a deep link that opens a specific destination inside the app on iOS 18 and later, can be assigned to keywords so it appears in search results for those terms, and can be used in Apple Ads. For a commerce app this is a direct way to map store pages to categories: a page for fashion, a page for home, a page for the current sale.

Custom product pages are not a replacement for A/B testing your default page, and the distinction matters for how you read results. Admiral Media explains the difference in why Apple custom product pages are not A/B tests.

Case study: ImmoScout24 matched pages to buyer and renter intent

Admiral Media worked with ImmoScout24, Germany’s leading real estate marketplace app, to build custom product pages for its two core audiences, renters and buyers, and tested them in Apple Search Ads Generic campaigns against search terms that matched each audience’s intent. Over four weeks in August 2022, in the relevant Generic ad groups:

• Buyers: CPA fell from €2.41 to €1.36 and cost per tap fell from €1.93 to €1.13.
• Renters: CPA of €0.80 versus €2.15 on the default page, and cost per tap fell from €2.00 to €0.69.
• Renters: tap-through rate of 28% versus 22% on the default page.
• Headline results: -43% CPA, -41% CPT and +17% TTR.

The mechanism is intent alignment. A renter searching for an apartment to rent who lands on screenshots about rental listings sees immediate confirmation that the app solves the problem, so more of them tap and install. The same principle applies to any commerce app with distinct shopper missions.

Deep links carry the intent past the install

A custom product page earns the install. A deep link protects it. If an ad promised winter jackets and the app opens on a generic home screen, the shopper has to search again, and many will not. Deep links send new and existing users directly to the relevant screen. For Google’s App campaigns for engagement, Google’s documentation notes that the app developer should complete deep link setup before deep links are added to campaigns, and that on Android 12 and later an incorrectly verified App Link sends clicks to the mobile website instead. Test every deep link on real devices before launch. For a primer on the setup, see Admiral Media’s guide to deep linking.

Building the channel mix for an e-commerce app

The best channel mix for an e-commerce app combines one or two scaled prospecting channels for new shoppers, a high-intent search channel for shoppers already looking, and a re-engagement layer that drives repeat orders. Which channels fill each role depends on the audience, the market and the product catalog.

Admiral Media does not prescribe a fixed mix, because the right answer depends on who the shopper is. The teen gifting app that won on Snapchat would have been a poor fit for a B2B office supplies app. The decision matrix below summarizes how the Admiral Media team typically evaluates each channel for commerce apps.

For deeper channel guidance, Admiral Media has published dedicated playbooks on Facebook ads for mobile apps and Google App Campaigns.

When awareness campaigns outperform user acquisition campaigns

Awareness campaigns can outperform user acquisition campaigns for commerce apps when personalized targeting is restricted or when the audience is young and socially driven, because cheaper reach builds demand that converts later through organic and search. This is counterintuitive, but measurable.

Admiral Media worked with Ønsk, Norway’s wishlist and gifting app developed with VG Lab, to grow among teenagers aged 13 to 17. EU rules prohibit personalized advertising to teenagers, which removes the targeting levers performance marketers normally rely on. The Admiral Media team owned creative production, including UGC and native assets, ran multi-platform testing across Snapchat, TikTok and Meta, and used awareness campaigns as a performance lever. Snapchat was the top performer for engagement and efficiency. Compared with user acquisition campaigns, the published results for awareness campaigns were:

• +60.3% day 7 ROAS and +41.2% day 30 ROAS: awareness campaigns returned more revenue per euro spent.
• 31% lower CPI: awareness campaigns also drove cheaper installs.
• 3.8x cheaper CPM: reach cost a fraction of the user acquisition campaigns.
• #1 Top Free App: Ønsk reached the top of the Norwegian free app ranking.

The lesson is not that every commerce app should shift budget to awareness. It is that the campaign objective with the best reported in-platform metrics is not always the one with the best business outcome. When targeting is constrained, cheaper reach plus strong native creative can generate demand that flows into installs and purchases through channels the awareness campaign never gets credit for. Measuring that requires cohort ROAS and incrementality thinking, which the measurement section covers.

Turning first-time buyers into repeat customers

Repeat purchases are where e-commerce apps make their money, so retention work belongs inside the user acquisition plan rather than after it. The most effective levers are deep-linked re-engagement, owned channels such as push and email, and catalog-based retargeting of lapsed buyers.

A commerce app that acquires a first order at break-even and never sees a second order has bought revenue at zero margin. The same app that converts a meaningful share of first-time buyers into second orders turns every acquisition cohort into a profit center. That is why the Admiral Media First-Order Funnel Framework ends on the second order, not the first.

The repeat-order playbook

• Onboard toward the first order, not the first session: shorten the path from install to checkout, prefill what you can and remove account creation friction where the business allows. Admiral Media’s guide to app onboarding optimization covers the patterns.
• Trigger the second order inside a defined window: use push and email around the category’s natural purchase cycle, weekly for groceries, seasonal for fashion. See Admiral Media’s push notification strategy guide.
• Retarget with products, not banners: catalog and dynamic creative show lapsed buyers the categories they browsed, as the South African delivery app results above illustrate.
• Deep link every re-engagement touch: a push about a price drop should open the product, not the home screen.
• Use discounts surgically: reserve incentives for lapsed buyers who show a pattern of returning with an offer, and measure the margin cost against the incremental orders.

Paid re-engagement changed considerably after Apple’s App Tracking Transparency limited user-level targeting on iOS. Admiral Media explains which approaches still work in app retargeting after ATT.

Measuring e-commerce app marketing correctly

E-commerce app marketing should be measured on cohort profitability, meaning margin-adjusted revenue from first and repeat orders per acquisition cohort, not on CPI or first-order ROAS alone. The metrics that look best in an ad platform dashboard are often the ones least connected to profit.

Measurement in commerce apps faces two compounding challenges. First, iOS privacy frameworks such as SKAdNetwork and its successor AdAttributionKit return aggregated, delayed and coarse-grained conversion data, which makes it hard to see repeat purchases at campaign level. Second, purchase behavior unfolds over weeks, so short attribution windows undercount the value of good cohorts. Admiral Media’s AdAttributionKit measurement playbook covers how to configure conversion values for commerce events.

The metrics that matter, in order

The Ønsk results show why tracking both day 7 and day 30 ROAS matters: the relative uplift of awareness campaigns was larger at day 7 than at day 30, which tells you how the value of each cohort develops over time. Admiral Media’s guide to incrementality testing for mobile apps explains how to check whether a channel is creating orders or just claiming them.

Margin-adjust your value signals

If you bid on purchase value, consider passing margin-adjusted values rather than raw GMV to the ad platforms, so the algorithm learns to find profitable shoppers rather than big spenders on low-margin items. The mechanism is simple: value-based bidding optimizes toward whatever number you send it. Send GMV, and it maximizes GMV. Send contribution margin, and it maximizes profit. The tradeoff is fewer high-value signals and potential volatility, so this works best for apps with stable margin structures across categories. Admiral Media covers value-based bidding further in predictive LTV bidding.

The most expensive mistakes in e-commerce app marketing

The most expensive mistakes in e-commerce app marketing come from optimizing the wrong event, reading promotional cohorts as normal cohorts, and treating the app as a copy of the website. Each one can make a scaling account look healthy while it loses money.

1. Optimizing toward installs for too long. The quick-commerce results above show the gap: installs up 64%, purchases up 1,210% once optimization moved deeper. Every week spent on install optimization after first-order volume is available trains the algorithm on the wrong user.
2. Judging cohorts during a promotion. Discount-driven cohorts look excellent on first-order metrics and often decay after the promotion ends. Read them on repeat orders at full price.
3. Running one store page for every audience. ImmoScout24’s renter CPA dropped from €2.15 on the default page to €0.80 on an intent-matched page. One generic listing leaves that efficiency on the table.
4. Sending ad clicks to the home screen. Without deep links, the intent created by the ad is lost at the moment of highest interest.
5. Briefing brand creative for performance channels. kaufDA’s creator ads drove 731% more user growth than regular ads. Polished brand assets rarely match native creator content on social platforms.
6. Measuring on GMV instead of margin. GMV-based ROAS can rise while the business loses money on every order. Finance and UA need to agree on the value signal.
7. Ignoring the app versus web split. Many commerce brands sell on both. If app campaigns cannibalize shoppers who would have bought on the web anyway, the incremental value is lower than the dashboard suggests.

How Admiral Media runs e-commerce app growth

Admiral Media runs e-commerce app growth as a single system covering media buying, creative production, store conversion and measurement, with the first and second order as the shared goal across every channel. The Admiral Media team has applied this approach to deals, quick-commerce, food delivery, gifting and marketplace apps.

In practice an engagement follows the First-Order Funnel Framework. The Admiral Media team starts with unit economics and event instrumentation, launches on the channels that best fit the shopper, and moves bidding deeper as conversion volume builds. Creative runs on a fixed testing cadence, with creator content, product-led video and dynamic creative tested side by side. Store conversion work runs in parallel, with custom product pages mapped to audiences and keywords. Measurement ties it together with cohort ROAS, second-order rates and incrementality checks.

The common thread across the case studies in this guide is that results came from moving the optimization target closer to revenue. The quick-commerce client cut paid CAC by 91% by moving from installs to first orders. kaufDA grew users by 731% by matching creative to how real shoppers talk. ImmoScout24 cut buyer CPA from €2.41 to €1.36 by matching store pages to intent. Ønsk lifted day 7 ROAS by 60.3% by choosing the campaign objective that fit its audience. If you want the same approach applied to your commerce app, the Admiral Media mobile app marketing team can review your funnel.

Frequently Asked Questions

What is e-commerce app marketing?

E-commerce app marketing is the work of acquiring, activating and retaining shoppers for a mobile commerce app, measured on orders and repeat purchases rather than installs. It combines paid user acquisition on channels such as Meta, TikTok, Google App Campaigns and Apple Search Ads with creative production, app store conversion, deep linking and lifecycle marketing. The goal is a profitable cohort of buyers who order more than once. Installs are a means to that end, not the result.

Should a shopping app optimize campaigns for installs or purchases?

A shopping app should optimize for purchases as soon as the campaign generates enough purchase events for the ad platform to learn, and use installs or registrations only as a starting point. Install optimization finds the cheapest people to install, who are rarely the most likely to buy. In Admiral Media’s work with a UK quick-commerce grocery app, moving optimization from installs to sign-ups to first orders contributed to a 91% reduction in paid CAC while paid purchases grew by 1,210%. The deeper the event, the closer the algorithm gets to your real customers.

What is a good CPI for an e-commerce app?

There is no single good CPI for an e-commerce app, because the right number depends on order value, margin, purchase frequency and market. A higher CPI can be profitable if those users order often, and a low CPI can lose money if they never order. Admiral Media recommends setting a maximum cost per first order based on contribution margin and expected repeat orders, then treating CPI as a diagnostic for creative and store conversion. Judge channels on cohort profitability rather than install cost.

Which ad channels work best for e-commerce apps?

The best ad channels for e-commerce apps depend on the shopper, but most successful mixes combine a scaled prospecting channel such as Meta or TikTok, a high-intent channel such as Apple Search Ads, and a re-engagement layer for repeat orders. Younger audiences often respond well to TikTok and Snapchat, while broad consumer categories tend to scale on Meta and Google App Campaigns. Admiral Media’s work with the Norwegian gifting app Ønsk found Snapchat to be the top performer for engagement and efficiency with a teen audience. Test channels against cohort ROAS rather than installs before committing budget.

Do custom product pages help e-commerce apps?

Custom product pages help e-commerce apps by showing each audience the products or use cases it is looking for, which raises conversion and lowers acquisition costs. Apple allows up to 70 custom product pages, each with its own screenshots, promotional text, app previews, optional keywords and optional deep link. In Admiral Media’s work with ImmoScout24, an intent-matched page for renters achieved a CPA of €0.80 versus €2.15 on the default page in Apple Search Ads. Commerce apps with distinct categories or shopper missions are strong candidates.

How do you increase repeat purchases in a shopping app?

You increase repeat purchases in a shopping app by shortening the path to the first order, triggering the second order within the category’s natural purchase cycle, and retargeting lapsed buyers with relevant products through deep-linked messages. Push notifications and email are the lowest-cost channels for this, while catalog and dynamic creative work well for paid re-engagement. Discounts should be used selectively and measured against the incremental orders they generate. The second-order rate is the clearest early indicator of long-term customer value.

How should e-commerce apps measure marketing success after iOS privacy changes?

E-commerce apps should measure marketing success on cohort-level revenue and margin, using day 7 and day 30 ROAS, second-order rates and incrementality tests, rather than relying on user-level attribution alone. On iOS, SKAdNetwork and AdAttributionKit return aggregated and delayed conversion data, so conversion values need to be mapped carefully to commerce events such as first purchase and order value. Incrementality testing helps confirm whether a channel creates orders or only claims them. Finance and marketing should agree on a margin-based definition of success.