Privacy and GA4
The last five years have been eventful in terms of privacy. A few countries within the EU have even stated that the industry standard tool, Google Analytics, is illegal under GDPR. I haven't seen any concrete or detailed backup plans yet, and frankly, I don't think we'll see one either. It’s months until Google plans to sunset Universal Analytics. However, they did recently push the end date for the 360 version of the tracking tool. Similarly, they also did the same for using third-party cookies in Chrome. It does give us a little bit of time to reflect and find new compliant solutions. My perception is that not many people are eager to switch to GA4, but soon it is a reality we need to face. It is quite the change and humans don’t always like changes. Given all circumstances, you are likely more susceptible now to explore other analytics tools as well which was not the case one year ago. As mentioned before I haven’t seen any concrete backup plans to GA, but the search trend for alternative analytics tools, such as Matomo is increasing. It is never bad to be aware of other solutions in case there would be a worst case scenario. Personally, I am relaxed and trust the transatlantic data privacy framework agreement, which was recently signed by the president of the United States and is expected to come into force six months from now. Read more about it here.

With GA4 comes a lot of new exciting features and reports. Machine learning is one of the highlights. Google says that the lack of tracking possibilities requires the need to extrapolate data.  

Media Mix Modelling (MMM)
Eventually, third party cookies will also die and we’d need to find new ways to measure our marketing activities. How do we know what's working and what data can we trust? It’s not a new question that we’re asking ourselves, but it’s more difficult to answer these days. One alternative is turning to Media Mix Modelling, often shortened as MMM. A MMM study allows businesses to evaluate and improve marketing activities without interfering with the integrity of user data.

The idea with an MMM study is to understand how each channel affects the total results, in order to get as accurate data as possible you either need a lot of data, or quite some variance in the data in order to isolate the impact from each channel. A MMM study will likely help you to argue for upper funnel activities as click based channels tend to steal most of the credit in different attribution models. MMM is usually associated with complexity, high costs, being time consuming and requiring a lot of historical structured data. It’s partly true, it is complex, but it doesn’t need to be expensive and time consuming. There are also workarounds for the historical data. There are many amazing, easy and user-friendly MMM tools. A big advantage is that the data flows on a daily basis which allows the model to be more accurate and you can work with the results on an ongoing basis instead of making it a one-time project which may be problematic as you may add or remove channels from the channel mix and the strategy in each channel may change over time. 

So if you haven’t looked into MMM yet, I highly advise you to explore and read more about it. Me or any of my colleagues are happy to chat about it over a coffee exchanging knowledge.
/Haresh