Which networks should an indie developer activate first? Start with two demand partners, then add a third only after measuring the initial eCPM lift. A second network usually creates a noticeable lift within the first week, while each addition produces less benefit by the eighth or ninth network.
Early additions matter because they increase auction pressure. One network bidding on your inventory is not an auction. Two create a real auction, with each aware of the other. A third adds a meaningful bidder when the first two have limited campaign budgets that day. After five or six quality partners with strong audience geo coverage, gains become smaller.
What a Demand Partner Actually Contributes
Every demand partner adds fill capacity and pricing competition to mediation. Fill capacity means campaigns target your app category, geo, and format. Pricing competition means those campaigns bid against other networks, raising the clearing price for your inventory.
A network may fill 95% of requests yet never bid above $0.50 on a slot where other partners regularly bid $1.50. That makes it a fill contributor, not a yield contributor. Both matter, but confusing them can make you dependent on networks that preserve fill while holding down eCPM.
Start with a few networks that overlap strongly on your main geo and format, allowing genuine competition for the same impressions. Three networks targeting the same user segment and country can compete harder than eight divided across geos and ad formats.
Evaluation Criteria: Questions Before Adding a Network
When screening a new demand partner, begin with four questions:
- Does the network have active campaigns in your main user countries? A network built around North American demand offers limited value when 70% of your impressions come from Southeast Asia.
- Does it support the specific formats you are running? Rewarded video and interstitial support are not universal; some networks focus on banners and will not place competitive bids on rewarded inventory.
- Does it have minimum traffic requirements before activating an account? Some larger networks set impression thresholds, and at 80,000 impressions per month you may not qualify for certain partners yet.
- The fourth consideration is easy to underestimate. A mediation adapter adds a few hundred kilobytes and initializes through the central mediation SDK. The same network's standalone SDK may add 2 to 5 megabytes and need a separate initialization sequence. With three or four networks, that difference matters on older devices and where storage is constrained.
APAC Demand Landscape Considerations
For apps with substantial audiences in South Korea, Japan, Southeast Asia, or other APAC markets, selection must reflect regional demand patterns. Some globally prominent networks have thinner campaign density in particular APAC markets, resulting in lower fill rates and CPMs there.
APAC publisher eCPM can vary substantially by country. A network strong on Korean traffic may have few campaigns for Thai or Indonesian users. Covering your top two or three geos by impression volume works better than pursuing global coverage through networks that are thin in your actual markets.
That is why geo-aware floor pricing matters in APAC. The proper floor for South Korean rewarded video impressions differs from the proper floor for Vietnamese banner impressions. One global average floor either forfeits revenue in high-CPM markets or reduces fill in lower-CPM ones.
The Diminishing Returns Curve in Practice
Adrop's early-access cohort aggregate data shows the general eCPM lift pattern as partners are added. Moving from one to two networks usually creates the largest lift, shifting from no competition to real competition. Two to three generally adds another meaningful lift through a multi-bidder auction on most impressions. Gains continue from three to five but compress. Beyond five quality partners with strong geo overlap, extra yield is often marginal.
For an indie developer, getting the first three to five right matters more than maximizing the count. Quickly activating eight networks, including several with weak coverage for your users, can hurt results. Poor-fill networks cause auction timeouts and may add latency without supplying competitive bids.
A Practical Starting Stack
For an indie app whose main audience is in APAC and whose formats include banner, interstitial, and rewarded video, start with three networks: one with strong pan-APAC rewarded video coverage, one with strong regional banner and interstitial fill, and one focused on your highest-CPM geo. Run this setup for three to four weeks, review winners by format and country, then decide whether a fourth partner adds competition or noise.
Track win rate, not per-network fill rate: the share of auctions each network wins and the price it wins at. A network winning 15% at high CPM is more valuable than one winning 60% at low CPM. Without price context, fill rate reveals little about yield contribution.
More networks are not inherently worse. Network count without geo fit and format coverage is simply a poor proxy for yield optimization. A focused stack of three well-matched partners in your core market will typically beat ten mismatched ones. Start focused, then expand from observed win rates and CPM deltas rather than assuming a larger stack is better.