Tech Mentat

Question everything. It’s a cliché that has been built into head of anyone who had decently intelligent parents or mentors. However, no matter how many times this has been drilled into your head – life happens. More poignantly, in the course of business, politics happens. When politics happen, no matter how many times we have been told or reminded ourselves to question everything, there is a tendency to forget. So here is a reminder:

Question Everything!

Since joining Healthline, I have noticed during many conversations and meetings decisions being made on gut feeling, past experiences or perceived optics (from users/investors) rather than data and tested learning. I don’t mean to paint the picture that all decisions at Healthline are made by the gut but I will say that I witness it more now than in past companies. Often when these decisions impact marketing, I raise my voice and ask for data or the need to set up a multivariate test to prove the worth of the idea. Yet, there are some decisions that were made prior to my arrival - the effects of these decisions can’t be fully known by the newer employee (they have no baseline or benchmark) so the line is towed and the button keeps getting pushed. I know I am preaching to the quire but let me repeat,

Question Everything!

For any company that drives significant amount of revenue from AdSense control, visibility (into what [keywords/sources] drive revenue), ad [placement] testing and the type of ads you define/allow are of the utmost concern. These things should never be taken for granted and every bit of it should be tested and proven over and over again to provide the best return [on investment]. As much as is possible, AdSense revenue should be tracked, treated and analyzed as if it were a physical product being sold.

For instance, while content-rich landing pages may provide the best initial quality score and have the highest perceived user experience, they might not be the best revenue generators OR provide the best user experience (as determined by positive feedback, registration percent & time spent on site). Content rich pages provide information but may not provide options (or multiple path availability) – similar to how Google is now refining searches, why not let your landing page mimic that idea? Likewise, while most people will instinctively drive a query on Concept A to a landing page containing information on Concept A. However, if your company has a well developed taxonomy such that you are able to find the relation between concepts – why not send the query on Concept A to a landing page for Concept B? Bad user experience? Let your users and revenue decide - if you aren’t giving them what they want, you’ll find out quicker than you want.

Recently, my challenge/question of the status quo revolved around the location and ad type (size) for both our SERP and article pages. The location and sizes had been initially placed based on “web standards.” But web standards are just that – standards or averages – they don’t necessarily work for every website. While specific numbers are, obviously, proprietary information I will say that the questioning resulted in an AdSense CTR increase of over 2%, over $2 increase in eCPM and over $2K daily increase in revenue. Point is, if you haven’t been doing so lately,

Question Everything!

For further reading on AdSense Optimization check out these pros.