Impact of Adjusting the Conversion Window on the Conversion Optimizer
Has anyone experimented with adjusting the conversion window to a shorter duration after having it set to 30 days and monitored its impact on the conversion optimizer? I am wondering if reducing the conversion window to 14 days instead of 30 will keep the conversion optimizer more focused on delivering at my CPA since the look back period will be shorter thus reducing the amount of poorly converting clicks that tend to occur between 20 to 30 days within the look back window after considerable spend has been established. The CPA is not influenced as much by poorly converting clicks when the spend is high so the average CPA is more difficult to increase or decrease. This gives optimizer the ability to deliver non converting clicks while still meeting the desired CPA goal. I'm looking to eliminate that portion of the spend and thinking a 14 day look back will reduce the amount of spend used in the CPA calculation thus offering less opportunity for optimizer to deliver poorly converting clicks since the CPA will be more sensitive to bad days where clicks are not converting. Has anyone experimented with this and what was the impact on spend, CPA, conversion volume, etc.
Re: Impact of Adjusting the Conversion Window on the Conversion Optimi
Hi @Chris M I have to admit, it's the first time I've seen a question about reducing the conversion window... In all the cases I've seen, the usual advice is to increase the conversion window since those conversions that come in late are usually just as valuable as those that happen "instantly".
A lot will depend upon how your CPC compares with your (net) conversion value. When conversions come in later in the path, they tend to also include multiple visits and often those multiple visits can include 2 or more paid clicks. If your CPC is a "trivial" amount when compared to the value of a conversion, I'd still argue that these later conversions are just as valuable as those that happen very quickly but if your CPC is a goodly percentage of your profit, you may well want to exclude those multiple click conversions.
However, you'll need to do some detailed analysis of how your later conversions come about, and why. The "how" is important because it could be the later conversions only require one paid click (with other visits being organic or direct), so they're no more expensive than the quick ones. The "why" is important because it could be that the time delay is reflected in the nature of the product/service, and/or influenced by external events such as weather, news, day of the week/month, etc. For example, I have a couple of clients that sell very expensive equipment that is typically bought by organisations rather than individuals so it's common for the buying process to be extended - they have to find what they want, then seek budget approval, get the budget signed off and so on. For one of these clients, over 40% of their conversions happen more than 3 weeks after first click, and it's completely understandable. I have other clients where it's normal for conversions to happen almost instantly, or within hours.
In summary, I'd say the first question to answer is why these later conversions are considered less valuable and, if there's a good reason, are there other settings/approaches that can address the problem(s).