- MMP
- App marketing
- Attribution
MMP: Do you need one at your current scale?
MMPs are often described as a must for app marketing. Here are our criteria for when to adopt one based on users, revenue, and ad spend, and how to use it so it pays for itself.
Dongle Media Lab8 min read
If you run an app, you'll hear that you have to use an MMP (a mobile attribution tool). But when you go to set one up, it isn't cheap, and it's not clear how it differs from the GA4 and ad platform reports you already have. The short answer: an MMP isn't something every app needs by default. It's a tool you need once you start spending meaningful budgets across multiple ad channels.
The problem an MMP actually solves
If you only look at ad platform reports, the numbers are always inflated. Meta says it drove the install, and Google claims the same result. Add them up and you get more than your actual installs. An MMP acts as the referee: it deduplicates these claims under one standard and decides who actually brought each user in.
- Deduplicates results across channels—credit is assigned under a single standard
- Tracks past the install—you only see real performance when you follow users through sign-ups, purchases, and return visits
- Deep linking—sends users from the ad to the exact screen in your app, reducing drop-off
- iOS support—handles privacy frameworks like SKAdNetwork for you
When to adopt one: three signals
It's hard to draw a hard numerical line, but the industry rule of thumb is fairly clear. An MMP is a tool for when you're running paid user acquisition (UA) at meaningful scale across multiple ad networks. If two or more of the following apply, it's time.
- 2 or more channels—if you only use Meta, there's not much to deduplicate. Once you go past 3, reconciling the numbers by hand becomes practically impossible.
- Monthly ad spend of several million KRW or more—this is the point where money wasted on wrong calls outweighs the cost of the tool.
- Revenue comes from what users do after installing—if the install isn't the end goal, as with subscriptions, in-app purchases, or repeat purchases, you need to see what happens next.
There are also clear signs that it's too early. If you barely advertise or use only one channel, most of your traffic is organic or viral, and you get no more than a few thousand installs a month, your money is better spent on the product and the message than on an MMP.
How much does it cost?
Most MMPs charge by monthly active users (MAU) or by the number of attributed conversions. Smaller volumes get a free tier (roughly up to a few thousand a month). Above that, pricing starts at around $0.05–0.07 per conversion and drops to $0.02–0.05 as you scale. For a mid-sized app, annual contracts in the range of several million to tens of millions of KRW are common.
Per-install pricing adds up faster than you'd expect. That's especially true if most of your installs never turn into revenue.
So before you adopt one, there's one thing to calculate: how many percent better do your ad spend decisions need to get for the tool to pay for itself? If your monthly ad spend is KRW 10 million and the tool costs KRW 500,000 a month, a 5% improvement in your decisions covers the cost. If your ad spend is KRW 1 million, the same tool becomes a burden.
Once you have one, here's how to make it pay off
Many teams set up an MMP and then look only at install counts. That turns it into an expensive install counter. If you've set one up, configure at least these three things.
- Define key events—use completed sign-ups, first purchases, and repeat purchases as your benchmark events, not installs.
- Connect deep links—send people from the ad straight to the relevant product or screen. Dropping them on the home screen throws away a click you paid for.
- Look at cohorts—compare channels not by “how many installs this month” but by “how much users who came in August spent 3 months later.”
If you want to go a step further, look at incrementality. It's a way to check whether users would have come anyway if the ad had been turned off, and it's especially useful in areas where results tend to be over-credited, like retargeting and branded search.
What an MMP won't do for you
An MMP tells you exactly where users came from, but not why they don't buy. Whether the message is weak, the product screen is unhelpful, or the price is off is still for people to judge. If you've added a measurement tool and results haven't changed, the problem isn't measurement. It's further upstream.
Dongle Media Lab doesn't start by telling you to add a tool. We first diagnose whether you need that tool at your current scale. If you do, we make event design, deep links, and cohort reports actually work. If you don't, we'll tell you to put that budget toward something more urgent.
