The Adwords Conversion Optimiser Algorithm Explained

In SEO, The Digital Marketing Blog by EloiLeave a Comment

Ah good times. Now Kelvin is back from a well deserved Honeymoon in Hawaii, we are going to be able to continue providing you with double blog love.

On one side, the best SEO and Social Media Marketing techniques are analysed & explained by Kelvin, and on the other, I will continue providing insights into setting up and running great PPC campaigns.

So here is my last post before Christmas, as I am leaving for my hometown of Paris tomorrow morning.

This week, Google’s Stuart Small dropped a bomb by saying at the great B2B marketing debate that 80% of searchers click on organic results and only 20% click on paid search results. If you work in Search Marketing, you have known this for a long time (or you should!) but the element of surprise comes from the fact that it is a Google rep has released what would be for them commercially sensitive information that could potentially cost them money as they do not profit from organic results, but instead make money by selling links.

The Gooogle description of the conversion optimiser via flickr


What does this tell us? Not much apart from the fact that if your PPC campgin is not tidy, well organised and performing well, you will not be able to get any enough conversions from that 20% of searchers to generate a great ROI.

So here is where my Google tool analysis will come in handy. If you are looking for conversions as your main KPI (key performance indicator) you must have tried using, or at least have heard of the ‘conversion optimiser’. It’s that tool Google put in place to allow you to indicate a maximum amount for each conversion (aka Pay Per Action, aka Cost Per Acquisition).

Great idea, because if I sell Easter chocolate bunnies, I do not want conversions to cost more than the production cost of the actual bunny…

Ah… How?

If you have tried implementing the conversion optimiser on y0ur campaign, you know a few things: firstly, you can only activate the conversion optimiser at the campaign level. Second, your campaigns need to have generated 300 conversions in the past 30 days in order to be ‘eligible’ for the conversion optimiser. Finally, if you have implemented it, you know it works… but do you know how it works?

The Conversion Optimiser is a combination of existing Google tools that you use on a daily basis (or should!) to which is applied an algorithm that will decide whether to display your ad.
“Do you mean that it will reduce the amount of impressions I get??”
“-Yes, but impressions are not what you are after, you are looking for conversions remember?” A lot of people believe that simply increasing impression will increase see clicks & conversions increase, but that actually is a gross misallocation of funds.

In order to define the best way of spending your money and displaying your ads, the conversion optimiser needs quite a lot of data (a bit like a marketer) which is why there is that annoying 300 conversions per month threshold. But if you do not have this amount conversions, you can use the tools that form the conversion optimiser’s algo to get the same results. The only thing you ll have to dedicate is a bit more time to it. But I have done it, and it can be done fast, even on large campaigns:

Variable A: Conversion Rate

First of all, the conversion optimiser looks at your best performing keywords (in terms of conversions of course). the [tag]conversion rate[tag] is then combined with the next variables to tell [tag]Google[tag] what to do and when. This is not a Google tool as such, but is the primary indicator of a good performing keyword.

Variable B: Position Preference

The position preference tool is great as well, as it allows you to choose which position you want your ads to be displayed in – but actually what it tells Google the opposite: where you do not want your ads to be displayed (ie under the fold –> low CTR=low quality score). furthermore, it is up to you to monitor which positions convert the best. If you are using this tool on its own, then my advice is to change positions every week, as usually you will convert at different rates on different days of the week. And check your bids, as they still determine whether you are bidding enough to appear above competitors.

Back to the [tag]conversion optimiser[tag], [tag]position preference[tag] data is used by Google to estimate which positions convert the best. That’s simple enough.

Variable C: Day Parting

Again, you can implement some day parting on your campaign without having 300 [tag]conversions[tag]/day, just make sure you have enough data to be sure about what times you convert better. For those who do not know about [tag]day parting[tag], basically it allows you to have your ads display at certain times, and not at other. (if you only deliver pizzas between 3 and 6 pm for example).

The conversion optimiser gets data [tag]AdWords[tag] and analyses what times of the day you convert better. Makes sense.

Variables A+B+C = [tag]Conversion Optimiser Algorithm[tag]

Now I am not an engineer, or an intense maths head but here is my interpretation of the way the Google conversion optimiser works. These three factors are interlinked, and I would not know the weightings of each, but they all play a significant part in the conversion optimiser algorithm:

Keyword Conversion Rate =< (Day Parting Conversion Rate + Position Pref. Conversion Rate) / 2

Then, it is just an automated task for Google to display the ads only when and where they convert better…

To finish, I would like to wish a merry Christmas to all our beloved readers and subscribers, ans which them the best of luck for 2008 if we do not talk to them before.


  1. Hey Kelvin – thanks very much for the kind mention. Drop by and tell us what you think of my wild stabs in the dark for 2008? dc

  2. Always keep an eye out for what my old employers at emap are up to David!

    You’re doing a great job presenting emap to the outside world. And from the sight of things you’re doing a pretty good job of moving the debate forward internally too.

    That’s what a great corprate blog does and slowly but surely it alters the make up of a company.

  3. “Never ban a website” – think this depends on internal culture and job role. If you have staff that post every 3 mins for 3 hours and should be selling products, think you are pretty justified at banning the site.

    think for sme the rules are slightly different! But some good points made!

  4. Your team addicted to Facebook Matt?

    You’re right that it can get in the way of work some time, but its an issue of trust in both directions.

    Let them know they’re allowed to use the site but don’t take the micky. Whatever you do though I wouldn’t just let the senior guys and marketing team have access that’s not going to do morale much good!

Leave a Comment