Diligence checklists for ad-network acquisitions
What I look for in a target's deck before I look at anything else.
When a fund asks me to do operator-side diligence on an ad-network acquisition target, I read four things first, in this order, before opening any model the founders sent.
1. The supply-side concentration page
Buried somewhere in the deck, usually around page 14. It tells you what percentage of revenue comes from the top 1, top 5, and top 20 supply sources.
What you want: A roughly Pareto distribution. Top 5 sources at 30–50% of revenue, top 20 at 70–80%. Nothing dominating.
Red flags:
- Top single source > 25%. One contract renegotiation reprices the whole company.
- Top single source < 5%. Sounds nice, but usually means the network has no anchor relationships and is buying every impression at market price.
2. The fill-rate-by-vintage chart
The chart they didn’t include. Ask for it. Fill rate of each cohort of integrated apps, plotted by how long ago they integrated.
What you want: Roughly flat lines. Fill rates that hold up over time.
Red flag: Old vintages with collapsing fill rates. It means the network is rotating supply through a stable demand pool — fine in the quarter, fatal as a business. Demand is what they’re really selling; they’re just borrowing supply to dress it up.
3. The eCPM disclosure
Most decks blur this with averages. You want the median eCPM by ad format and by geo, week-over-week for the last 12 months.
What you want: Stability or slow growth.
Red flags:
- A step change. Means a single demand contract started or ended.
- Smooth growth that exactly matches a public benchmark (Adjust, AppsFlyer, etc). Means the chart was reconstructed from a benchmark, not measured. You’re not looking at their data.
4. The eligibility / brand-safety logic
The least sexy section. The one that produces 80% of the silent revenue leaks I’ve seen in this category.
What you want: A clear, version-controlled spec for which impressions are eligible for which campaigns. A team that knows where the logic lives and can show you the most recent change with a date attached.
Red flag: Anyone hand-waving about “our brand safety stack.” That phrase means it’s a Rube Goldberg machine of grandfathered rules and nobody on the engineering team can confidently tell you what each one does.
I’ve watched a $4M/year leak in this layer get caught by a single PR after six months of executives blaming bidder behavior. The team had the data the whole time. Nobody had read the eligibility logic in two years.
There’s a longer checklist. These four catch most of the deal-killers before you’ve spent an hour on the model. If a target survives them, you’re now doing diligence on the company, not on the deck.