Throughout recent years many clients of ours have reached us to discuss the business challenges they face and the first thing they rush to show us is the data analytics from the marketing efforts their former marketing agencies had achieved.

They focus on traffic acquisition, increased money spending, conversions, conversion value, declining ROAS (Return On Ad Spend) and a ton of other metrics, while no discussion has been preceded regarding the theoretical aspect as well as the technical setup of conversion tracking; if it reflects the business’ purposes and if it is correct and accurate to the extent possible.

In this article we will get to introduce you with the theoretical view of conversion tracking. What it is, what is its purpose, how we approach it.

Depending on the business you own, there are certain user actions that drive revenue in your business’ bank account. These actions might be a purchase if you own an online consumer goods store, an app install if you are active in the mobile app distribution sector, a successful form submission if you provide professional paid seminars and so it goes…

So, before starting online advertising, the first thing you need to clear out and define in detail is:

• what action you consider as a conversion,
• what value (in numerical format) is each conversion contributing to your business (for an e-shop owner, each purchase of product is a conversion and the product price is the value of each conversion, the conversion value),
• can this type of conversion be set up in the marketing platforms that will be used (Google Ads, Facebook Ads, Google Analytics, etc.) and, if yes, how.

Definitely no! Why? First we have to sort out the aforementioned key points. Because from a cost perspective, Google and all such corporations have their cost tracking in place and are ready to measure in great detail your marketing spending. And they will not give away a penny.

If you have not accurately set up your business’ conversion tracking, then the metrics you see might be greatly misleading. Let’s see a simple example.

You run an online hospitality business that gets paid via user redirects to your hotel partners’ websites. You get paid by a fixed amount per redirect. So the redirect is your conversion and the fixed amount per hotel partner is your conversion value.

Let’s say that your hotel partner named “Dream vacations” pays you €0.05 per redirect. The person that implemented the conversion tracking setup made the careless mistake of setting the conversion value of “Dream vacations” redirect to €0.50.

Lets get down to numbers with a certain user case.

When you open your Google Ads report you see that your campaigns show an average Cost per ad Click (avg. CPC) of €0.10 and for the last 30 days you had 100 clicks with 50 conversions, meaning a total spend of 100 X €0.10 = €10 and a total conversion value of 50 X €0.50 = €25. Wow, do you see the ROAS that the report is displaying? It is €25 / €10 = 2.5 or in percentage format 250%. It seems like you are profitable, since for every €1 you spend, you get back €2.5. Correct?

Unfortunately, no! Remember that the correct value per conversion is €0.05? Let’s do the calculations concerning the total conversion value one more time. The correct revenue amount would be 50 X €0.05 = €2.5. Oh, snap! This means that your actual correct ROAS was only €2.5 / €10 = 0.25 or 25%, meaning that for every €1 you spend, you get back only €0.25. Sad but true…

In Click-wise we always consider that conversion tracking is one of the first things to investigate. Our attack plan is relatively simple, but also very effective.

1. We need to understand your business and get to know your KPIs (Key Performance Indicators).
2. We define every significant conversion and its corresponding value.
3. If setup does not exist, we proceed in setting up conversion tracking on each platform.
4. If setup exists, we proceed in the next step.
5. We test conversion tracking using fake conversions that we have full control of. This way we see the flow of conversions and their value from our ad click to what is reported on each platform.
6. If exceptions arise, we come up with custom solutions.

Unfortunately,

a) not every type of conversion that we can think of can be set up and

b) we can never rest assured that even if the conversion tracking is correctly set up, it will also be extremely accurate.

This may be due to lack of necessary software, lack of technical know-how or time or even care required from people outside of our sphere of influence (i.e. from employees in a client’s store), etc.

For example, if the aforementioned client was getting paid not on a fixed amount per redirect, but depending on the vacation deal that the user would book in the partner’s website (i.e. a 20% percent on the booking the user makes), this means that the client cannot know beforehand the actual value per conversion (if there is no dynamic flow of information between the client’s website and the partner’s website, or in other words cross domain tracking), but can only make an estimation based on the previous month’s average value per conversion that is derived from a reconciliation of conversions and total value between the two sides (situation a).

Another example, if an employee of yours is making online purchases via a Google search paid advertising campaign of your store (because it appears in the top of Google search results and he/she is bored to scroll down and click on the organic result of your website or even better keep it on a bookmark for which the URL will contain certain tags that will specify that the traffic comes from an employee) for customers that have visited your physical store and has neglected to inform the marketing specialist, these purchases are pushed as conversions through the Google Search paid advertising channel, giving the misleading idea that the marketing efforts in Google Search paid advertising are successful in an extent that is above expected (situation b).

Or if purchases that are cancelled are not constantly dropped from the platforms data, this means that the revenue displayed in the platforms is overestimated (situation b).

When a technical permanent fix is not applicable, practical workarounds do the job.

A workaround on such issues is to observe the month over month mismatch of tracked conversion value versus the actual conversion value and if the mismatch is stable through time calculate a correction ratio, i.e. if in January tracked revenue is €1000 while actual revenue is €500 and in February the tracked revenue is €10,000 while actual revenue is €5,300 we can make the assumption that the correction ratio is approximately 0.5 or 50%, meaning that if in March the tracked revenue comes up €20,000 we will know that the correct value will be around €20,000 X 0.5 = €10,000.

Even when conversion tracking is set up in the best possible way, there will always be the question which conversion is attributed where, what time window we allow for a conversion to be recorded or how many conversions can be matched to the same user. What do we mean by all these? We will explain as clear as possible.

The potential converter (user that is about to convert) may have made a long customer journey till he/she completes the conversion. This journey may have began from a Display campaign that showed your brand and they clicked on but bounced at once from your website, then a Facebook remarketing ad banner showing up on the news feed, then a Google organic result that your store ranked in the second position till, finally, he/she, by remembering your brand, typed your store’s website URL directly in the browser bar, entered your website and completed the conversion.

Now, this conversion will appear in your reports to have happened via the Google Organic channel if the attribution model your marketing specialist uses is set to “Last non-direct click”. This conversion could be wholly attributed to Google Display channel if the in-use attribution model was “First click” and so it goes.

This is a setting that is user-specified and defines how long forth in time will a conversion action be recorded and matched to an ad click that is performed today.

For example, if today I search a keyword in Google search bar, a search ad pops up and I click on the ad, but I immediately bounce, or in other words leave the website, the marketer of that website would see in his/hers Google Ads user interface one impression and one click for that specific keyword for today’s date.

If after three days I decide to finally convert on that website and the marketer has chosen a conversion window of three days or more, then in the Google Ads user interface one conversion with its respective conversion value will be matched to the keyword I had searched. And it will be shown for the date that I had performed the ad click and not on the date I finally converted.

Sometimes a user may perform an ad click and continue with converting just once in the website. But what about websites that offer the choice of multiple conversions in a landing page. Such a website could be a travel meta search website, that displays offers for multiple flight companies and each click on an offer redirects the user and is counted as a conversion.

In such a case wouldn’t it be more correct for our conversion tracking setting to allow recording multiple conversions for a single ad click. The aforementioned user could have clicked on the ad and booked a ticket for himself/herself and another three for his/her friends. All these conversion counts mean revenue to the website and credit to the marketing channel.

All these are classic topics that are always discussed intensively among marketing specialists. Truth to be told, there is no one-solution-fits-all answer to which setting is best. The choice depends on your type of business and on the assumptions that the marketing specialist chooses to proceed with.

In the current article’s examples we mainly focused on conversion tracking for Google ads. However, the conversion tracking technical implementation of a business regards all in-use marketing platforms (Facebook ads, Google Analytics, etc.). Each platform has each own settings for the marketer to deal with. In another article we will reveal more about the technical aspect of conversion tracking and get to show more technical details about how conversion tracking can be set up.