Thank you for checking back in for my Arbitrage Diary – it really means a lot that you’ve been reading and following this story!

Dan’s smashing it in France this weekend with his other love – his band. Do I feel cheated? Not at all! Why? I get to bring you the blog!

I’m going to be talking about AdSense optimisations today, to keep us on track with our schedule of variables and the best ways to optimise them:

As Publishers Variable/Metric Aiming for?
AdSense/Content.Ad Widget Placement and Size (depends)
AdSense/Content.Ad Text or Image? Text/Border Colours (depends)

Guess what? AdSense testing is super easy. And SUPER worthwhile!

Google does all the hard work for you, so once you know why and how to get testing, it’s a case of sit back and enjoy (sort of)!


Why test?

AdSense testing is A/B testing of your adsense ad units, and can apply to really general variables (e.g. whether your ad unit serves up text ads or image ads), or to really, really specific variables (e.g. which shade of neon green text works best for your text-only ad units).

FYI, if you haven’t realised/weren’t aware (I hadn’t, before we started this project…), this is the difference between image and text ad units:


blog 10 pic 1


[Haha even now, with a screenshot on google docs, i’m avoiding accidentally clicking on them at all costs – kind of feels like Big Brother is watching…][Also, a later article will focus on arrows, how to live without them/ how to get them back/ our general experiences/ so stay tuned for that!]

Testing constantly creates improvements, and you can run tests non-stop. You can have one per unit running at all times – 3 units, max of 1 big one per page [though we are also just starting to experiment with adding link units https://support.google.com/adsense/answer/15817?hl=en-GB   – more details in a later blog!], so 3 tests on the go all the time. The variations of tests help reduce banner blindness, so in of itself the process is beneficial. The clear improvements and financial gains make it all the better!

Google takes the test you want to run, then runs equal amounts of your traffic to each of the two variables you are going to test (no chance of coordinating that without Google’s help!). Once you let it run for a few days, you’ll have clear-cut feedback as to what’s happening and why. All very easy to understand! The main metric success is measured in is $$$$, which is what we are here for, so it’s obvious from the get-go what is being gained.

In the couple of months we’ve been doing this, there has been no more immediate and verifiable a way to make effective gains to our site’s ability to earn than AdSense testing, and it should be exactly the same for you too. Other things can make bigger differences (content, traffic sources etc), but they take time, and some deep trawling of analytics. AdSense testing is quick and easy to implement, with clear % gains. It’s stupid to ignore – you’re leaving money on the table if you don’t start asap!) Big swing example:


blog 10 pic 3A


Actually, when we got started with our site, I spent almost all of my time pouring through the specific Google Analytics site pages to see our AdSense stats to scope out relevant Adsense stats like CTRs, revenue, ads clicked, impressions, and adsense ecpm, rather than looking at the AdSense site itself. However, it is full of very useful and specific info that can be drilled down into, so if you’ve neglected to snoop around the resources on there, do take the time to have a look at what’s what. Get on your AdSense page and get stuck in!


How to test

In Blog 5 we went through setting up AdSense units (https://www.nativeadbuzz.com/blog/setting-up-adsense-and-content-ad-adverts-on-your-site/), so you should be ready to roll with testing.

For today, what matters is the Optimization tab:


blog 10 pic 3B


You’ll first see Google’s suggestions for quick and easy experiments to run for your site, based on what the data your users is telling them. It’s an easy place to start!


blog 10 pic 4

  1. Under the Opportunities tab, choose Experiments:


blog 10 pic 5


  1. From there, you’ve got the option to test either Ad Unit Settings, or Allow/Block ads.

Ad unit settings is what we test almost exclusively – it is more ‘layout’ based, and we find it the easiest to start with. Allow/Block ads is more specific content-related stuff – i.e. if your site’s niche is about veganism, it would make sense to block ads from Fast Food restaurants etc… [we’ve only recently started running these tests ourselves, but seeing some positive improvements already – 11% increase in RPM by blocking ‘Home + Garden’ ads. Who knew…!]


blog 10 pic 6


  1. So, choose ‘Ad unit settings’. You’ll come to this screen. The AdSense business end:


blog 10 pic 7


  1. Pick which of your three units you’d like to test, name the test (we recommend an easily repeated naming system – you’ll be running a lot of tests so it’s good to be able to easily track

them all). Proof of pure obsession with testing? Here you go:

  1. Next, pick the experiment setting. For this example, i’m doing a Text ad style – a test on a text ad unit variating the colours and font. Click create, and you’re up and running. Simple as that!

Worth starting with are some of the tests we have had really good success with:

  • Testing the standard blue of AdSense font colour against the dominant colour of your site (for us, red).
  • image only vs text only, specifically on the middle ad unit.
  • Image + text vs text only, specifically on the bottom ad unit.
  • Verdana font vs standard AdSense font (Verdana is big and visible and seems to have a good positive effect on $$$ performance)

ALSO – sometimes, and incredibly counterintuitively, we have run a test for a week or so, gotten a certain result – case in point, when we’ve tested image vs text – and then rerun the same exact test, only to get a different result. There are lots of reasons why (running different campaigns, so different traffic sources and users / different time of the year so different ads being served, etc). The lesson is, ALWAYS test. There will ALWAYS be gains to be made!

  1. Google measure success with two different categories – Quality score, and Overall change…


blog 10 pic 9blog 10 pic 10


…basically, the better each does, the more $$$ your unit will make in terms of RPM after that change is made. Simples!

Here’s an example of a test that went very well, and made a big gain for us. While the $$$ amount is low (this was from when we were only spending $25 or so a day), the margin is massive – 54%.


blog 10 pic 11


That’s what this is all about – making gains in the percentages that will pay off as your spend scales up.

And just as importantly, here’s a test that ‘didn’t go well’. Which really just means that the new variation was not as strong as the original. But look how much better the original was! It’s essentially a process of elimination, so no harm, no foul!


blog 10 pic 12


Basically, get testing! AdSense is fully equipped with the tools to make serious gains with very minimal effort, and the information you will gain about your site and what best practices are will be super insightful. It’s often VERY counterintuitive, but for us that has just meant we’ve realised how important it is to trust the numbers. Intuition is important, but the results speak for themselves!

Dan is back from hurting French eardrums next week, so will be taking you through the next steps in our optimisation process.

Hit us up with any questions if you have any – a good place to start would be in the brand new Native Ad Buzz forum – http://forum.nativeadbuzz.com/ Some great stuff in there already! Otherwise, give us a shout in the comments.


Thanks again!