Last year I wrote an article called “6 predictions for the Analytics Industry in 2022“. With the year 2022 coming to a close, it’s time to evaluate how well these predictions did. As with every prediction, evaluation is needed to evaluate, learn and improve. Please note that this evaluation is biassed and subjective in some way. It’s based on a personal perception and experience of the market.
1. Event based tracking will be the primary way to track web behaviour
Summary / TLDR;
I predicted that event based tracking would become the primary way to track web behaviour as the main vendor - Google - would move to an event-based measurement model with GA4.
Evaluation:
The story didn’t exactly unfold as expected. Users are not as satisfied with GA4 as Google would have expected which makes brands evaluate and eventually migrate to different alternatives. This causes a split in tools based on measurement models. The “old” way of measuring - read GA3 - alternatives such as Piwik Pro or even more extreme forms of simplified measurement such as Plausible and, on the other side, the “event-based” way of measuring such as Mixpanel, Amplitude and Snowplow. At large I’d say there is definitely a shift therefore I’m inclined to sign-off on this predictions. Even though some brands hold fast to their desire to keep measuring the way they did in GA3, we see a lot of brands moving into event-based measurement. Mainly because of what you’ll read in the evaluation of prediction #4.
2. Web analytics will evolve in a multitude of fields tailored to the needs of different stakeholders
Summary / TLDR;
I predicted that analytics would split into 3 distinct fields being - marketing analytics / product analytics / customer analytics.
Evaluation:
I find this one personally hard to evaluate. Let me lay out the reasons why I’m not comfortable signing off on this one:
Why I might be wrong:
Product analytics companies are starting to do marketing analytics. See Amplitude’s (a product analytics tool) announcement on their development of marketing analytics capabilities. Knowing that there is a product analytics tool now embedding marketing analytics makes the distinction between the fields blurry.
Amplitude positioning as a CDP. Knowing that they’re a product analytics solution in the first place while adding customer data capabilities makes the distinction between customer analytics and product analytics blurry.
Customer data platforms are starting to embed customer journey analytics. See mParticle’s announcement on acquiring Indicative. Which brings them very close to doing product analytics as the measurement model is the same and the visualisation is already integrated.
Pure dashboarding tools are starting to integrate product analytics features such as funnel visualisation and sankey diagrams. This doesn’t necessarily mean that they’re user friendly, it just indicates that dashboarding and analysis capabilities are converging and that different types of data can be treated in a single location.
Why I might be right:
Some of the best solutions in their fields stay focused on what they do and are very successful at it. Examples are Mixpanel for product analytics and GA4 in marketing analytics.
If you spend millions in the Google space I still believe that GA4 will be the go-to place for you to do marketing analytics. Mostly because Google is able to offer modelling in GA4 as it’s integrated with the entire Google advertising space. With privacy measures increasing, ad blockers rising and browsers ditching 3rd party cookies Google advertisers will become more reliant on Google’s modelling, which is only available in GA4, going forward.
I’m unable to say where this will go as I feel the skill set and knowledge required from analysts to excel in the three respective fields goes beyond just technology capabilities. I’ll leave this on in the middle for now.
3. The role of the analytics team will change. From being a service desk to facilitating data democratisation
Summary / TLDR;
I predicted that analytics teams will move from being the centre of all analyses to the facilitator of data democratisation.
Evaluation:
At Human37 we’re seeing an increasing amount of missions to actually achieve this from an organisational and technological perspective. Obviously this is linked to prediction #4 (see next prediction) and the building of analytics as part of a larger stack. Where different analytics products are being used by different teams, each with a different focus. The analytics team currently stays at the centre of things and increasingly starts to support self-serve capabilities by helping to build out the proper infrastructure. I’d say that the shift is starting for some organisations and that this prediction is being realised as we speak.
4. Analytics is no longer a standalone solution, it’s part of a stack
Summary / TLDR;
I predicted that analytics would be part of a larger stack and that data would flow simultaneously to (or through) different systems based on a single measurement. Data points in analytics platforms would be the same as in other platforms (such as an event in your ESP) and therefore the data would be activatable/transferable from one platform to another. Data observed in your analytical solution would be activatable in other parts of your stack.
Evaluation:
I believe there’s enough proof that marketing technology and analytics are converging into a single stack. At Human37 we’ve been lucky enough to work with some awesome brands to actually help construct, optimise or evaluate some of these stacks. From an operational point-of-view we see more and more analytics platforms being a destination and therefore not being fed by non-native SDKs. Examples are Mixpanel and Amplitude implementations that are fueled using SDKs from a CDP such as Segment or mParticle or by rETL tools such as Census and Hightouch. Important to note - and coupling back to what was said in prediction 1 - is that all (or at least most) of these systems use event based measurement at the core. The more we’ll move into stacks, the more it will be event-driven and the more it becomes event-driven the easier it will be to be integrated into a stack. The same goes for implementations fed by a rETL deployment.
From a strategic point-of-view we see an increasing amount of “coupled” implementations from the get-go. Where the operational point-of-view tells the story on how brands fit a single solution as a destination into their stack we now see that brands are starting to look into buying different solutions at once and optimise integrations as of day 1.
From my point-of–view this prediction holds up.
5. The rise of a new KPI - the identification rate.
Summary / TLDR;
I predicted that marketing organisations will start using identification rate as a KPI. This means the fact that a user provides the brand with a lasting identifier such as email or phone number in order to start building a relationship. The identification rate as a KPI can be seen as the rate at which a company is successful at generating identified and opted-in first party data (aka zero-party data).
Evaluation:
I do believe that this one is a fairly accurate prediction. Not saying we’re currently living in a world where identification rate is a universal KPI, but definitely moving in that direction. Best example is Google’s Privacy for Agencies and Partners Certification where this was brought up.
6. As much as I would like to be wrong: Building dashboards or reports and looking at them will still be the primary way of “doing analytics” for the majority of companies in 2022.
Summary / TLDR;
What the title says. The end station for data is a report / dashboard, not an activation channel.
Evaluation:
Even though we’re seeing more and more data activation use cases (see prediction #4) we still see an awful lot of data being “stuck” in reports and dashboards. That’s mostly the case in organisations of larger sizes or corporate environments where data and marketing stacks aren’t front and centre. Therefore I believe this prediction will always be a relevant one to make. Even if it’s just to challenge those organisations that are report/dashboard richt but data activation poor.
Curious to hear about your evaluation. If you’re interested in any of these topics or if you have questions, don’t hesitate to reach out.
If you need help with regards to any of those topics, feel free to reach out.
Comments