A Simple Guide to Analyzing Paid Traffic and Avoiding Fraud

EN
What tools can you use to analyse traffic sources and detect fraudulent users?
A Simple Guide to Analyzing Paid Traffic and Avoiding Fraud
Published
22.02.2023
|
devtodev

Selling and purchasing user traffic is a big business. As with any other business, it has fraudsters roaming around and wanting to sell you a bad product for the price of a good one. In this article, we are going to talk about purchasing traffic for your app or game and ways to avoid being scammed by shady traffic providers.

As a matter of fact, the only way to succeed in this kind of business is to take full control of it. How can you do it? Obviously, by using analytics! Even a basic inspection will allow you to find sources that provide you with fraudulent users. After you find it, you can simply get rid of it. 

Let’s take a look at how you can analyze your paid (or sourced) traffic.

Read more: Accurate LTV Prediction using Machine Learning Model

Metrics

The best way to inspect your traffic is by using metrics. There are some metrics that you can use on the first day of paid acquisition (“instant-action metrics”). Other metrics need to acquire some data before you will be able to use them for data evaluation (let’s call them “delayed-action metrics”).

Instant-action metrics are:

  • Tutorial completion

  • Registration conversion rate

  • Install conversion rate

  • Level 1 / Level N completion

  • View N items in the shop or M sections of the website

  • Or any other useful action that the users can perform on the first day

Delayed-action metrics are:

  • Retention rate (D1, D2, D7, etc.)

  • Number of sessions per user

  • Purchase conversion rate

  • N day cumulative ARPU

  • N day ROI

  • LTV (LTV > 3CPI)

Paid traffic analysis pie chart
Screenshot from devtodev demo

You need to set a certain time limit for these metrics. For example, as a rule of thumb, your organic users make their purchases during the first week. In this case, you can compare the purchase conversion rate of your organic traffic and the traffic you have bought. Or you can compare the average session number and/or length of your users acquired from different paid sources. Or with your best traffic performance. Read more about the benchmarks below.

Read more: How to Integrate an Analytics System into your Game

Target events

To see how good the acquired users are, you can analyze the first actions of the users — do they complete the target actions they are supposed to complete? Do they do something else that will later lead them to completing those actions?

How do you take a peek into what the users do during the first day(s) after they have installed your app or game? You can simply create several target events and monitor them using an analytics platform. The actions are similar to the “instant-action” metrics we have listed above — item views, first purchases, completion of the first level(s), etc. However, you can turn any user action into a tracked metric, e.g. shop open, steps of your marketing funnel, battle with a friend, etc. 

When you analyze target events, it’s very important to pay attention to the number of unique users because fraud users tend to repeat their actions (e.g. install the game again and again and complete the first level).

Read more: SQL for Beginners: How to Track First In-App Events

Paid sourced traffic analysis platform
Screenshot from devtodev demo

Analytics

You can use both metrics and target events to create cohorts and then compare them to each other and to the benchmarks.

To analyze the sourced traffic, you need to have some benchmarks. They are unique for each project, but the most widely used are retention, purchase conversion, cumulative ARPU, and ROI. You need to establish certain values for your key benchmarks (using organic traffic metrics or some reference sources) and compare them with the values of the traffic you purchase.

Read more: Best Game Analytics Platform: devtodev vs DeltaDNA

Sourced traffic cohort analysis
Screenshot from devtodev demo

Recommendations on Traffic Acquisition

devtodev recommends you use the following rules if you want to acquire good traffic:

  • Analyze users sourced from each particular campaign. Don’t think that if you acquired good ones from this particular channel last time, this time it will be the same good. Always analyze traffic sources, not the channel in general.

  • Exclude whales from the analysis. Whales are the users who spend tons of money on your product. Usually, there are very few whales in each game or app but they can contaminate the data and make it unreliable. For example, you may think highly of a particular traffic source because it contributes a lot of revenue. However, if you exclude this one whale from the analysis, you will find out that this user source is, in fact, borderline fraudulent.

  • Reports from paid traffic providers may have the “Other” column and up to 20-30% of users may come from that unknown source. You want to make sure that you know precisely what sources they used because only in this case you can analyze them.

  • You need to establish a routine for acquired user analysis. For example, monitor certain metrics every morning.

  • Diversify your sourced traffic and don’t buy it from a single provider — use at least two or three.

  • Work with verified partners (especially when you’ve just started). You can easily find a list on the Internet.

Read more: How to Retain Players in Mobile Games

Easy Algorithm for Working with a New Paid Traffic Provider

After you have chosen a new provider, you want to integrate it and see whether it’s worth the money. So, you need to:

  1. Slice and dice the traffic
  2. Analyze it
  3. Disable junk sources and keep the good ones (or simply stop working with the company)
  4. Integrate the next provider

This way, the quality of your paid traffic will improve and become more predictable.

Read more