How tracking works — from the marketing perspective
There are a lot of articles talking about web tracking and tracking tools like Google Analytics. The article shall give you some insights from the perspective of marketing.
Privacy vs. Conversion
The core target for online tracking tools is to assess the performance(i.e., the conversion rate) of different marketing channels. The math is easy, for customers associated with a channel, we get how many of them had performed a specific action(like registering), and the total number of customers, and therefore we have the conversion rate as the quotient of the two numbers.
But how to identify the marketing channel of a customer is the tricky part. The traditional way is to mark the customer browser session on the occasion of their first visit with some extra id that does not easily expire — the cookies — you get the cookie, and you know where they are from. Most of the tracking tools(including Google Analytics) are still using cookies today, but privacy concerns have already scheduled the retirement of cookies, and a lot of browsers and tools specifically forbid the use of cookies. Privacy regulations like GDPR are also unfriendly to cookies, especially the third-party cookies that Google Analytifs use.
With the warning of the retirement of cookies, some tools are adapting techniques like browser fingerprints. The fingerprints are calculated based on the users' browser version, OS version, and whether some Javascript features are supported … Tools declaring that they do not use cookies probably adopt the methods. The issue with the method is that most of the fingerprints are explicitly calculated, so they are either easily non-unique or likely changed due to user-side modifications. The numbers are not useful if you have a rather small customer base.
In the future, something like Google’s Privacy Sandbox is giving customers explicit control of how their data may be requested for ads. The conversion rate is more like to be an estimation rather than an accurate number since identifying a specific customer is even more different.
Sampling
If you have a rather large volume of customer data, the report on Google Analytics will be generated based on a sampled subset of your data. You may frown at the issue, but actually, I can guarantee that the sampling is OK for you.
Based on the trend of tracking, you will soon have no explicit data on your customers. Furthermore, if you have a large volume of data enough to be sampled, the sampled data should be of the same statistical distribution as the unsampled data.
But you do not have enough data, i.e., enough customers. Your data won’t be sampled.
Managed vs. Self-Hosted
Regulations like GDPR are preventing customer data from being transferred outside countries or Europe. Managed services like Google Analytics are not compatible with those regulations.
If you need strict compliance, self-hosted services are required, and you need to deploy and maintain those services yourself. That will bring you a raise in the cost of human resources and things like cloud services.
From the trend based on web tracking, I think the situation grow un unfriendly with small merchants. If you are considering a marketing strategy, maybe an omni channel strategy is better than optimizing a single channel.
