Updated: Apr 19


14 years ago, venture capitalist Dave McClure, coined the acronym AARRR - Acquisition, Activation, Retention, Revenue, Referral - , a model for tech startups to achieve maximum growth by understanding and optimizing key performance metrics at each stage of the customer journey. This model is widely accepted not only because it turns the spotlight on to 5 essential metrics to achieve growth but also due to the "MECE" approach of such metrics.

This post outlines the steps to adapt the AARRR model to a digital publisher's yield management strategy by tying mutually exclusive and collectively exhaustive events across different components of a publisher's tech stack.

Problem Statement.

1 Too many data sources. Traffic monetization involves multiple layers and multiple tech solutions: from driving and retaining users to bid stream optimization, data compliance, etc.

2 Isolated data. Tracking and analyzing all metrics related to traffic monetization is often times an isolated and non-inclusive process due to the distinctive nature of the metrics and of the reporting systems involved.

3. Disconnected dots. Analysing metrics separately makes it very hard to demonstrate the impact of one layer on the other layers.

4. Low impact. Optimization efforts conducted on a specific layer will show little to no improvements on yield, unless the metrics of other layers having a direct or indirect impact on such layer are not optimized accordingly.

Solution & Methodology


Provide a MECE framework and set of metrics that can be used by AdOps and revenue managers to track and correlate the impact of metrics appearing to be unrelated to one another to achieve higher yield.


Design a yield management model - inspired on the AARRR model- where KPIs from the following 5 "monetization stages" are mutually exclusive and collectively exhaustive:

1. Acquisition: Stands for driving the highest possible number of users to your site or app (traffic).

2. Activation: Stands for turning the highest possible number of unique users into into ad impressions (ad inventory).

3. Retention: Stands for keeping the highest possible UX level despite advertising related activity (ad quality).

4. Revenue: Stands for selling the highest possible number of ad impressions (sales-thru-rate).

5. RPM optimization: Stands for selling the highest possible number of ad impressions at the highest possible price.(yield optimization)

Process & Execution

  1. Get your ducks in a row. List all active components of your monetization stack.

  2. Build the funnel. Classify your active components under one of the 5 "monetization stages" above.

  3. Define your KPIs. Define at least 1 KPI/metric per monetization stage.

  4. Create the rules. In a waterfall system pair a KPI from a given monetization stage (hereinafter "Event") with a KPI from the next monetization stage (hereinafter "Benchmark").

  5. Activate the funnel. Create an alert triggered every time a fluctuation on the average last 7 days data for the Benchmark results on a X% increase or decrease of the Event's daily performance.

  6. Expand your metrics. Repeat step 5 adding more KPIs to each layer pairing them to new or existing KPIs from other layers.

Use case.

Expected outcome & validation.

  1. In the example above, we have chosen one metric for each stage of the ad monetization funnel (Total Sessions, Impressions, Viewability Rate, Bid Rate, ARPS).

  2. We have also pulled for each metric the average results of the last 7 days and used them as Benchmarks.

  3. Results from previous day (last 24 hours) show that all metrics have decreased, with exception of Total Sessions and Viewability Rate that experienced increases of 25% and 1% respectively.

  4. Under regular scrutiny, that means looking at isolated data, one could easily attribute the decrease of Revenue per Sessions to a slower reaction to return bids by the SSP mix in our network reflected by a decrease of the Bid Rate, which wouldn't be conclusive for an AdOps as the fluctuation was minimum. Further scrutiny would turn the eyes of the AdOps manager to traffic trying to tie the decrease of ad Inventory to a drop of daily visits to the page, just to find that traffic actually increased.

  5. By creating alerts at each stage of the monetization funnel an AdOps can easily identify that the main issue driving revenue down is at the ACTIVATION level, caused by the inability to turn traffic into ad inventory thus focusing the attention to page layout or ad units related issues.

  6. As a result, the AdOps has not only identified the anomaly causing the drop of ARPS but has also done that effortless and immediately.

Useful Resources.

- Use yieldPass Smart Dashboards to unify all your data sources in one place.

- Use yieldPass Surveillance Software to set customized alerts and set your AARRR model.

50 views0 comments
Banner 300x250.png

Recent Blog Posts

Banner 300x250.png