Quick answer: How much an MMP should cost depends less on the rate you negotiate than on the pricing model you sign. Per-conversion pricing scales with how many installs and events you buy, spend-based pricing scales with your budget, and flat tiers stay fixed until you cross a volume cap – so the same app can pay very different shares of spend on each.
There is no independent benchmark that says an MMP should cost a set percent of ad spend. Vendor rules of thumb exist, but they are vendor claims. The useful test is your own: convert every quote into a share of your planned spend, at your real cost per install, at the spend levels you expect to hit over the contract.
The pricing model, not the discount you negotiate, decides what your MMP costs you at scale. Pick the model that fits your unit economics first, then negotiate the rate.
What do MMPs publish about their pricing?
Most mobile measurement partners publish a free tier and a self-serve tier, then move larger advertisers to custom enterprise pricing. The public pages are worth reading closely, because they show which unit each vendor bills on – and that unit is what drives your cost at scale.
AppsFlyer
AppsFlyer's pricing page lists Zero, Growth and Enterprise plans, with 12K conversions free for the first year on Zero and Growth and $0.07 per conversion after that on the pay-as-you-go Growth plan. Enterprise is custom pricing. AppsFlyer defines a conversion as an install or user action that your marketing drove and that AppsFlyer measured, and says organic installs it does not attribute are free – per the same AppsFlyer pricing FAQ. It also flags the Zero plan as not suitable for apps with paid activity.
Singular
Singular's pricing page lists a Free plan that includes 15,000 paid conversions, a Growth plan at $0.05 per conversion and a custom Enterprise plan. The same page tells buyers they can pay by ad spend or by conversions, which makes Singular one of the few vendors that names a spend-based option publicly. It does not publish the spend-based rate.
Kochava
Kochava's pricing page is closer to a subscription model: a free plan that includes 10K conversions a month, a Foundation plan starting at $500 a month that includes 33K conversions a month, and an Enterprise plan starting at $2,000 a month that includes 200K MAU a month. Kochava says Enterprise can be billed on attributed conversion volume or on monthly active users.
Adjust and Branch
Adjust bills paid usage on attributions – its help center notes that TV commercial attributions are billed as regular attributions. We could not load a current Adjust price sheet when this guide was drafted, so we have not stated its tier limits or rates. Ask for them in writing.
Branch's pricing page names three tiers – Basics, Essentials and Enterprise – and gates the higher two behind a demo request. It publishes partner limits per tier, not per-unit prices.
Published list rates are a starting point. What matters is the billing unit, what counts toward it and what happens past a threshold.
Roughly what percent of ad spend should an MMP cost?
There is no reliable industry percentage. We could not find a named, independent survey that measures MMP fees as a share of ad spend; the percentages that circulate come from MMP vendors' own blogs, and one widely repeated rule of thumb comes from a vendor arguing for its own model. Treat any single number as a sales claim, not a benchmark.
What you can do is calculate it. For per-conversion pricing, your MMP cost as a share of spend equals the per-conversion rate divided by your cost per billable conversion. That one ratio explains most of the confusion around this question – two advertisers on the same contract can pay very different shares of spend because their cost per install differs.
Run it at your real blended cost per billable conversion. If the result is a meaningful share of the budget that buys the installs, the model is wrong for your economics – and a better rate on the wrong model only shrinks the problem.
How do per-conversion, flat and spend-based models compare?
Each model shifts risk to a different party. The table summarises how they behave.
| Model | What you pay for | Cost grows with | Works best for | Watch out for |
|---|---|---|---|---|
| Per conversion (attribution) | Each attributed install or measured action | Volume of paid conversions | Apps with moderate or high cost per install | Low-CPI traffic, re-engagement and event billing inflate the count |
| Flat or tiered subscription | A monthly or annual fee with an included volume | Tier jumps and overage once you pass the cap | Stable, predictable volume | Paying for headroom you never use; steep overage rates |
| Percent of ad spend | A share of the media you run through the MMP | Your budget, regardless of volume | Low-CPI apps with high install counts | Paying more when CPIs rise, even though measurement work did not change |
| MAU-based | Monthly active users | Your retained audience | Apps where engagement, not paid volume, is the main cost driver | Growth in organic users raising your bill |
Per-conversion pricing
This is the default for most published self-serve plans. It is easy to understand and cheap when volume is low. The catch is that your bill tracks paid volume, not value. If you move budget into a cheaper source – rewarded traffic, broad paid social, re-engagement – your installs rise faster than your spend, and so does your MMP bill.
Flat and tiered subscriptions
A fixed fee with an included allowance makes budgeting simple and rewards efficient growth inside the tier. The risk sits at the edges: a small app pays a high share of spend for capacity it does not use, and a fast-growing app can hit an overage rate or forced tier upgrade mid-contract.
Percent of ad spend
Spend-based pricing ties the fee to budget, not volume. That suits cheap, plentiful installs. It is a poor fit when your cost per install is high or rising, because you pay more for the same measured installs – and the vendor earns more when your media gets more expensive.
What does each model cost at different spend levels?
The worked example below uses hypothetical, illustrative numbers – not any vendor's quote and not our network data. It assumes one app with a $4 blended cost per billable conversion, a per-conversion rate equal to AppsFlyer's published Growth rate, a hypothetical 2% spend-based rate, and a hypothetical flat plan of $3,000 a month including 200,000 conversions with $0.03 overage. Swap in your own figures.
| Monthly ad spend (hypothetical) | Billable conversions at $4 | Per conversion at $0.07 | Spend-based at 2% | Flat $3,000 + overage |
|---|---|---|---|---|
| $20,000 | 5,000 | $350 (1.75%) | $400 (2%) | $3,000 (15%) |
| $100,000 | 25,000 | $1,750 (1.75%) | $2,000 (2%) | $3,000 (3%) |
| $500,000 | 125,000 | $8,750 (1.75%) | $10,000 (2%) | $3,000 (0.6%) |
| $2,000,000 | 500,000 | $35,000 (1.75%) | $40,000 (2%) | $12,000 (0.6%) |
Three things fall out of this table.
- Per-conversion and spend-based pricing both hold a constant share of spend at a fixed CPI. They only diverge from each other when your cost per install changes.
- Flat pricing is the most expensive model at low spend and the cheapest at high spend. The crossover in this example sits between the second and third rows. Below it, a flat plan is paying for headroom.
- Overage terms decide whether flat pricing stays cheap. In the last row, overage makes up most of the bill. A worse overage rate would erase the advantage.
Why does your cost per install decide the winner?
Per-conversion and spend-based pricing break even at one cost per billable conversion: the per-conversion rate divided by the spend-based rate. Above that CPI, per-conversion pricing is cheaper. Below it, spend-based pricing is cheaper. At the hypothetical rates above, the break-even point is $3.50 per conversion.
Here is the same hypothetical $100,000 month at three different costs per billable conversion – again, illustrative numbers only.
| Cost per billable conversion (hypothetical) | Billable conversions | Per conversion at $0.07 | Spend-based at 2% |
|---|---|---|---|
| $1 | 100,000 | $7,000 (7%) | $2,000 (2%) |
| $4 | 25,000 | $1,750 (1.75%) | $2,000 (2%) |
| $15 | 6,667 | $467 (0.47%) | $2,000 (2%) |
This is why the percent-of-spend question has no single answer. An app buying cheap installs – rewarded, hyper-casual or broad social – can pay several times more on a per-conversion contract than a subscription or fintech app paying a high CPI on the same contract. If you are adding rewarded campaigns, our guide to rewarded app CPI rates covers what those campaigns pay, which is the input you need here.
Two practical points follow:
- Count billable conversions, not just installs. If re-engagements and in-app events count toward the bill, your cost per billable conversion is lower than your CPI, which pushes per-conversion pricing up.
- Model next year's mix, not this year's. If you plan to shift budget toward cheaper sources, a per-conversion contract that looks efficient today can get expensive mid-term.
What should you negotiate before you sign?
Negotiate the model first and the rate second. Use this checklist on every quote and renewal.
Billing unit and definitions
- What exactly counts as a billable conversion – installs only, or re-attributions, re-engagements and in-app events too?
- Are organic installs free? Are SKAdNetwork postbacks, web-to-app or CTV attributions billed, and at what rate?
- Does the price change by platform or region?
Volume and thresholds
- What volume is included, and what is the overage rate past it?
- Are tiers based on monthly or annual volume, and can unused volume roll over?
- Is there a minimum commitment, and what happens if you fall short?
Model flexibility
- Can you switch between per-conversion, flat and spend-based billing at renewal – or mid-term if your mix changes?
- If spend-based, is the rate tiered down as spend rises, and what counts as spend (all media, or only media routed through the MMP)?
Add-ons and hidden costs
- Which features sit outside the base price – raw data exports, fraud prevention, audiences, incrementality, cost aggregation, SSO, extra seats?
- Do any ad partner integrations carry a fee? Kochava's pricing page, for example, says integrations are typically free but notes exceptions for some ad tech vendors.
Contract terms
- How long is the term, what is the auto-renewal notice period, and is there a price cap on renewal?
- How long is raw data retained, and can you export your full history if you leave?
- What support level is included at your tier?
Run every answer back through your own break-even sheet at three spend levels: where you are, where you plan to be and a downside case. The cheapest quote at today's spend is often not the cheapest contract over its full term.
If you are planning next year's acquisition budget and want to test a performance network channel alongside your existing sources, talk to our advertiser team about a capped test you can measure in your own MMP.
FAQ
How much should an MMP cost? It depends on the pricing model and your cost per install more than on the headline rate. Convert each quote into a share of planned ad spend at your real cost per billable conversion, at current and expected spend levels. If the share is small and the data drives budget decisions, the price is defensible. If not, change the model before you negotiate the rate.
Roughly what percent of ad spend should an MMP cost us? There is no independent benchmark. The percentages in circulation come from MMP vendor blogs, not named industry surveys. Calculate your own: on per-conversion pricing, divide the per-conversion rate by your cost per billable conversion. Cheap installs push that share up; expensive installs push it down. Compare that result against a spend-based or flat quote at the same spend.
Is per-conversion or percent-of-spend pricing cheaper? Neither is always cheaper. The two break even at one cost per billable conversion – the per-conversion rate divided by the spend-based rate. If your cost per billable conversion sits above that point, per-conversion pricing is cheaper. If it sits below, spend-based pricing usually wins. Low-CPI apps tend to favour spend-based or flat pricing.
Do organic installs count toward MMP billing? On some plans, no. AppsFlyer's pricing FAQ says installs and actions it does not attribute to your marketing are not counted as conversions and are free. Other vendors define billable units differently, and some bill on monthly active users, which includes organic users. Get the definition of a billable unit in writing before you sign.
Can a small app use a free MMP plan? Several vendors publish free tiers with a conversion allowance, but read the limits. AppsFlyer's pricing page says its free Zero plan is not suitable for apps with paid activity, while Singular and Kochava publish free plans with included conversions. A free tier works for early testing; plan for the paid model you will move into as spend grows.
When should we renegotiate our MMP contract? Renegotiate when your traffic mix changes enough to move your cost per billable conversion, when you are approaching a tier cap, or ahead of annual budget planning. Bring a break-even sheet showing your cost under each model at expected spend. A model switch at renewal often saves more than a rate discount on the existing model.