Advice about ironSource ads best practices tends to arrive as a long checklist where every item is weighted equally, which is the opposite of how the results actually distribute. In app user acquisition, a small number of decisions carry most of the outcome and the rest are hygiene. Knowing which is which saves months. The channel lives inside mobile app inventory, largely games and utility apps, and the mechanics of that environment shape everything about how a campaign should be built.
This is written for teams running app growth rather than for people evaluating the platform in the abstract. We will stay away from specific rates and placement inventory, both because they change and because the numbers that matter are the ones from your own account. The approach below is what we apply when we take over an account that has been spending without a clear view of what is working.
Creative carries more of the result than targeting
In this environment the creative is the targeting. Users are interrupted mid session inside another app, so the asset has to earn attention rather than receive it. Playables and short video that show actual usage in the opening seconds consistently outperform polished pieces that build slowly toward a reveal. Anything that looks like a trailer competes with entertainment; anything that looks like the experience itself competes with curiosity, which is a much easier fight.
Treat creative production as a continuous process with a fixed weekly or biweekly output, not as a project that concludes at launch. Concepts wear out faster here than in most channels because frequency accumulates quickly. Keep proven assets running while new concepts are tested alongside them, and record which underlying idea worked rather than just which file won, so the learning survives the asset.
Measure past the install
An install is an intermediate step, not an outcome, and optimising toward install cost alone reliably produces volume that never returns. Decide which downstream signal represents value for your app, whether that is retention on a given day, a tutorial completed, a subscription started or revenue within the first week, and hold every source to that standard. The gap between two traffic sources with identical install costs and very different retention curves is the entire game.
That requires measurement that actually works. Verify your attribution setup with test events before scaling, confirm that in app events are passing back correctly, and check that your event definitions mean the same thing in the platform as they do in your own analytics. A surprising share of accounts we inherit are optimising against an event that stopped firing correctly after an app update and nobody noticed because the dashboard still showed a number.
- Establish a downstream quality metric before launch and refuse to judge sources on install cost alone.
- Test one clear creative concept at a time with enough budget to reach significance.
- Review source level performance regularly and exclude publishers that deliver volume without retention.
- Keep campaign structure simple so budget concentrates and optimisation can learn.
- Re verify event tracking after every app release.
ironSource ads best practices for bidding and structure
Fragmentation is the most common structural error. Splitting budget across many small campaigns feels like control but starves each one of the data it needs, leaving everything permanently in a learning state. Consolidate around a few campaigns with meaningfully different purposes, give each enough budget to accumulate signal, and make changes one at a time so you can attribute the effect.
On bidding, decide whether you are buying to a target cost or buying volume, and be consistent about it. Large sudden changes to bids or budgets reset optimisation and often change the composition of traffic in ways that are not visible until the retention data arrives a week later. Incremental adjustments with an observation window between them are slower and considerably cheaper than the alternative.
Source quality management is ongoing work
Not all inventory within a network behaves the same way, and a portion of it will deliver installs that never open the app a second time. Review performance at source level on a regular cadence and build exclusion lists based on your own downstream data rather than reputation or assumption. This is unglamorous maintenance that pays for itself repeatedly, and it is the step most often skipped when a team is busy.
Watch for patterns rather than single bad days. A source with poor retention across several weeks is a decision; a source with one weak week may simply have had different placement luck. Keep the record long enough to tell those apart.
Where the app itself is the constraint
No acquisition work fixes a first session that loses people. If retention is weak across every source, the honest read is that the problem sits in onboarding, load time or the early experience rather than in the media buy. Spending more in that state simply pours new users into the same gap. We would rather pause scaling, fix the first five minutes of the app, and restart against a product that holds the users it receives.
The short version of ironSource ads best practices: produce creative continuously, measure past the install, keep structure concentrated, manage source quality as routine work, and be willing to stop when the constraint is the product. Most of the rest is detail that matters only once those are in place.
Keep reading: Ironsource Ads · Advertising