Yesterday we reviewed some of the changes in the ways that people spend their marketing budgets. Today we’re going to talk about why user retention metrics, uninstall rates, and growth in fraudulent activity warrant your attention.
As an app marketer in 2016, you are right to be at least somewhat concerned about what I call user decay (app abandons and uninstalls) as well as app fraud.
At Apsalar we do spend significant time analyzing fraud and unusual activity in the app world. I do want to draw a distinction between fraud and unusual activity. Fraud is a deliberate attempt to mislead and cost advertisers and consumers money, whereas unusual activity encompasses the unexpected – things that intuitively don’t seem normal, but aren’t the result of a deliberate intent to mislead. But taken together, fraud and sources of unusual activity have a big effect on marketing performance and ROI. Let’s take a moment to examine how clicks are affected.
We sometimes ask marketers to estimate how many clicks it takes, on average to drive an install. In an ideal world, that number would be one. Someone would click and then complete a download/install. Most people guess somewhere between 2-10. I am sure that at least some of you will be surprised to hear that, when we conducted our last analysis, the global figure was 83. 83 clicks per install.
And the US is only a little bit better than average. What causes all this?
- Some people click and then decide not to download.
- Sometimes fat fingers mistakenly drive up click counts.
- Many networks continue to rely on CPC deals with individual publishers in order to deliver on their CPA contracts with brands.
- As marketing spending in our category increases, the payoffs for fraudulent activity have grown markedly. That attracts more fraudsters and may increase the overall rate of fraud in the industry.
There are other factors as well. But whatever the reasons, it’s clear that you shouldn’t rely on click counts to optimize your business if you are being measured by revenue.
Now let’s talk about uninstalls. As you may know, Apsalar was the first attribution company to launch uninstall attribution measurement. Using Apsalar Uninstall Attribution, you can compare campaigns, creatives, regions, cohorts and vendors to see which drive the lowest uninstall rates. Uninstalls can be a great diagnostic tool for understanding why one tactic drives better ROI than another. Shortly we will release to the public our category level findings, but I can reveal that the range of uninstall rates varies hugely based on vendor, campaign, region, etc. That is one key reason why optimizing to CPI can lead to misallocations – all installs are not of the same quality.
Better performers tend to be large, premium providers and networks, providing some evidence that you get what you pay for. Media that demonstrate affinity or intent. So vertical networks or companies that are good at targeting like-minded users can perform well here. Nonincentivized tends to outperform incentivized.
Finally, we found significant disparities between North American uninstall rates and those found in Latin America and Asia, which likely relate to the relative maturity of markets and the average memory size on devices in those regions. When a phone has a small memory, people have to uninstall apps that they might like perfectly well in order to add apps for which they have an urgent need or desire.
The point to all this discussion is that unusual and fraudulent activity warrant your attention, because they may have significant implications on your optimization processes. It’s best to optimize to your KPIs rather than to surrogates.
Tomorrow: Proactively Improving Post-Install Customer Relationships.