Mobile ad fraud can be solved with smarter team incentives

This guest post is written by Shamanth Rao, a seasoned mobile user acquisition executive and a popular contributor to the Mobile Dev Memo slack team.

In the wake of Uber’s recent bombshell fraud lawsuit against a number of mobile ad networks, one question arose within professional mobile marketing circles: did the folks at Uber really have no idea what was going on? Some of the ad networks named in the lawsuit are fairly commonly known as bad actors — and most kinds of fraud can be detected easily.

Did the folks at Uber not know about this fraud? Or did they just not care?

I will hypothesize that at least a part of the answer lies in how most user acquisition managers are typically evaluated — and thus compensated. User acquisition managers’ incentives can often explain why they might turn a blind eye to fraud, willingly or otherwise.

How so? Most user acquisition managers and teams are typically assessed based on:

  1. Performance of paid installs (only). What happens to overall installs (organic + paid) matters far less to user acquisition managers;
  2. Upstream ROAS targets (most commonly on D7 ROAS, which is oftentimes taken as a leading indicator of lifetime ROAS). What happens past D7 matters far less for many teams.

The consequence of the above is that many user acquisition teams are not incentivized to care about what happens beyond their immediate mandate of maximizing Day 7 ROAS from paid traffic. The financial health of the app or the company is hardly something that they are driven to care about.

This means that if a fraudster is engaging in click fraud to ‘steal’ organics, well, that doesn’t hurt a UA manager’s paid traffic goals (as long as these installs meet performance benchmarks) because this traffic shows up as paid traffic. Similarly, if a fraudster is simulating a first purchase event, it isn’t going to hurt a UA manager’s metrics because their Day 7 revenue and ROAS will continue to look good in their MMP and internal systems (even if the ‘purchases’ are not real).

What is a marketer or executive to do then?

In a system where user acquisition managers have limited incentives to address fraud, other steps are necessary to correct this gap with (at least partially) misaligned incentives. These 3 strategies can go a long way:

1. Make at least a part of UA managers’ evaluation based on overall ROAS, not just that of paid traffic.

Yes, this will make it so that UA managers may get rewarded for organic growth — and it may be argued that this is for work that they haven’t done. That said, this overall ROAS is the metric that matters most to the company’s financials, since this signifies dollars in and dollars out.

With an incentive plan like this, UA managers get compensated for their contribution to the overall ROAS — so if the company’s overall financials are hurt by fraud, UA managers have an incentive to fight it.

2. Backtest past ROAS performance and have some portion of UA managers’ assessment contingent on this.

Once you calculate ROAS after Day 7 and see what happens at D14 and D30, you get a clearer picture of what is happening after the first monetization event. You can still use the immediate feedback loop of D3 or D7 numbers for immediate optimization and look at downstream metrics to understand true performance. Indeed, the best teams look very closely at actual financials from older cohorts.

In fact, in the first episode of the How Things Grow podcast, Adam Lovallo speaks of the first sign he saw that the daily deal space might be slowing down — this happened when his team at LivingSocial backtested projected revenues from a few months prior, and found that the actuals were nowhere close to projections. Admittedly very little of this lag was on account of fraud, but this backtesting gave them a sign that all wasn’t well with their financial prospects.

A similar approach to backtesting can often reveal whether the D7 ROAS continues to be an indicator of long-term financial health of a user acquisition team’s work.

3. Make at least a part of UA managers’ evaluation contingent on overall revenue & DAU growth, not just ROAS. 

If a UA manager is doing her job well, the app or product is meeting its growth goals and the revenue and DAU are growing (or if these metrics were declining, their slide is arrested). If a UA manager is incentivized for the product’s growth goals, they are incentivized to fight fraud, since non-trivial amounts of fraud will dent this growth trajectory. 

Indeed, one company I know tracks cost-per-DAU-increment as a KPI for its UA team, and I know of others that incentivize UA teams in part on overall revenue and profitability goals. Having UA managers be incentivized to maximize DAU and revenue growth (not just ROAS growth) can be a very powerful way to incentivize them to truly fight fraud.

At the end of the day, UA is a financial function as much as it is a marketing one, and aligning UA managers’ incentives with the financial health of the product and company oftentimes can help mitigate problems of misaligned incentives that can often result in UA teams turning a blind eye to fraud.

Shamanth Rao is the founder of the mobile user acquisition agency RocketShip HQ, the host of the How Things Grow podcast, and a regular contributor to the Mobile Dev Memo Slack community. Shamanth previously served as the VP of Growth and User Acquisition at FreshPlanet and worked on the Growth team at Zynga’s NYC studio.

Photo by Maarten van den Heuvel on Unsplash