Xiaomi Ads metrics that matter are the ones tied to what happens after the click: cost per install, cost per in-app action, day-1 and day-7 retention, and return on ad spend from your mobile measurement partner. Impressions and click-through rate on Xiaomi placements such as GetApps, the system browser and lock screen help diagnose creative, but they do not prove a campaign is working.
Why does Xiaomi need its own measurement approach?
Xiaomi Ads, sold through the Xiaomi Global Ads platform, reaches users inside the phone’s own system apps: the GetApps store, the Mi Browser, the themes app, the security and cleaner apps, and the lock screen. These placements behave differently from social feeds. A user may install an app straight from a preloaded store recommendation without ever visiting a landing page.
That means the classic web funnel of impression, click, page view and conversion does not fit. Measurement must start at the install and continue inside the app, which puts your mobile measurement partner (MMP), such as AppsFlyer, Adjust or Singular, at the center of reporting.
The core metrics, from first impression to revenue
Group metrics by the question each one answers. That keeps dashboards readable and stops teams from optimizing a number that only explains part of the story.
- Delivery: impressions, reach, frequency by placement.
- Interest: click-through rate, install rate from store page views.
- Acquisition cost: cost per install (CPI), cost per registration, cost per first purchase.
- Quality: day-1, day-7 and day-30 retention, sessions per user, tutorial completion.
- Value: revenue per user, ROAS at day 7 and day 30, payback period.
- Integrity: install-to-open rate, click-to-install time, share of rejected installs.
Cost per install is a starting point
CPI on system placements can look attractive compared with other channels. That is exactly why it needs context. A cheap install from a user who never opens the app costs more per active user than an expensive install from someone who returns every day. Always calculate cost per retained user alongside CPI.
Retention shows whether placements fit your app
Compare retention cohorts by placement. Lock screen and cleaner app placements reach broad audiences, while GetApps search and category placements reach users already looking for apps. In a typical gaming account, for example, store search traffic might retain better than lock screen traffic even at a higher CPI. Your own data decides; do not assume.
How do you connect Xiaomi Ads to an MMP?
Integration happens in two places. In your MMP, enable Xiaomi as an integrated partner and generate tracking links with the required macros. In the Xiaomi Global Ads dashboard, attach those links to campaigns and confirm postbacks for installs and in-app events.
- Map the in-app events you want to optimize toward, such as registration or purchase.
- Send event value and currency for revenue events.
- Check that the attribution windows in the MMP match your reporting rules.
- Run a test install on a Xiaomi device and confirm it appears in both systems.
Once postbacks work, optimization strategies that target in-app events become available. Without them, the platform can only optimize for installs, and you lose control over quality.
Which numbers will not match, and why?
Xiaomi’s dashboard and your MMP will rarely agree perfectly. The platform may count installs differently, use another time zone or include view-through attribution. Your MMP also deduplicates installs that several networks claim.
Treat the MMP as the source of truth for installs and revenue, and the ad platform as the source of truth for spend and delivery. Reconcile weekly. A gap that suddenly grows, for example from a few percent to a third, usually signals a broken link, a changed macro or a fraud issue rather than normal variation.
How do you spot invalid traffic?
Mobile ad fraud exists on every network. Your MMP’s fraud protection tools flag patterns such as click injection, click spamming and device farms. Learn the signals so you can question odd results quickly.
- Click-to-install time that is extremely short across many installs.
- High installs with near-zero day-1 retention from one sub-placement.
- Many installs from devices with identical models and OS versions in a short window.
- In-app events that occur in impossible sequences or speeds.
When you find such a pattern, pause the source, document it and request a review through your account contact. Keep records, since rejected installs often need evidence for refunds or credits.
Creative metrics belong in the same review. Compare click-through and install rates per creative and placement, since a banner that works in GetApps may perform poorly on the lock screen, where users see it in passing. Retire creatives whose install rate drops week over week, and keep a short log of what each variant tested so results build into knowledge.
Awareness campaigns need different metrics
Not every Xiaomi campaign is about installs. Brands sometimes use lock screen or theme placements for reach. For these, look at unique reach, frequency, video completion rate and brand search lift. A short brand lift survey or a geo test, where some regions see ads and others do not, gives a more honest answer than click-through rate.
Keep frequency in check. Lock screen impressions are highly visible, and showing the same creative too often can irritate users rather than build familiarity.
Building a dashboard leadership will trust
A useful weekly report fits on one screen. Put spend, installs, CPI and cost per retained user at the top. Below that, show ROAS at day 7 by country and placement, then retention cohorts. Add a short notes column explaining changes such as new creatives, budget shifts or paused sources.
Report trends over several weeks instead of reacting to single days. Mobile campaigns fluctuate with weekdays, holidays and store featuring. A seven-day rolling average smooths noise and makes decisions calmer.
Setting targets by country
Xiaomi’s user base is concentrated in markets such as India, Indonesia, Southeast Asia, parts of Europe and Latin America, and costs differ across them. Set CPI and ROAS targets per country based on your own revenue per user there, not a global average. A target that is right for one market can block profitable delivery in another.
Finally, agree on definitions before arguing about results. Decide whether an install counts at first open or at store download, which time zone the report uses and which attribution window applies. Many disputes between growth and finance teams disappear once these rules are written down.
Where to start this week
Confirm that your MMP sends at least one meaningful in-app event back to Xiaomi Global Ads, and test it on a real device. Then build a one-page report with cost per retained user and day-7 ROAS by placement. Pause the weakest placement after it has spent enough for a fair read, and move that budget to the one that keeps users the longest.
Keep reading: Xiaomi Ads · Advertising