Strategic analytics

I wrote a post recently on how to build a measurement strategy. And got a couple of messages from folks going, oh I should do that! And well experienced people too. And I think it’s an important topic, if the folks in charge of analytics don’t keep it linked to the business strategy, then naturally others won’t give it as much value.

And that’s how data & analytics teams end up backed in corners, just taking orders. Which is not a fun place at all. In that post, I suggest teams should create a data request brief, to help filter and get better requests upfront. Let me know what you think of the piece.

This POV is reflected in DBT’s 2024 State of Analytics Engineering report with only 14% of data professionals “strongly agree” that their organization sets clear goals for their team. I do think this is the wrong way around though, analytics team should set the goals, based on the business objectives.

Data quality really is a big issue - also touched on in the report. I remember chatting with the ex CIO of LinkedIn, and they were describing what it took to clean up all the user inputted data for analysis. This was some time ago, but the amount of effort it took, was monumental.

The tools now are better, but this still seems the most pressing issue. Getting the data, tidying it up, collecting the right data points. But the better data the better the output. Which is why all the AI companies are trying to lock down data sources and find new sources of data. WSJ covered this, story below.

I was reminded of this in reading Robyn’s documentation. Robyn is the open source MMM that Facebook released. Anecdotally people have mentioned to me that it favours Facebook. I mean that’s believable and a cheap punch. But if you dig into it, the model is going to better understand sources that give the best data. Chances are your Facebook ad data is better, than say your DOOH, or TV. Those with the best data win! This is going to be the same no matter which model you use, really the challenge is to find, grow & maintain high quality data for the models.

The other tangent I touch on in the measurement piece, is preparing for AI but also design. I think more analytics teams should get comfortable with hiring designers, to help come in and show different ways to visualize the data. To make sure the end output, is prepared in a box with a nice bow. Presentation matters.

Notable stories this week

• MeasureCamp New York is on this weekend.

• Marketers find new ways to buy programmatic ads without relying on user data.

• DBT’s 2024 State of Analytics Engineering is out.

• Google gives Universal Analytics shutdown 1 July deadline.

• For Data-Guzzling AI Companies, the Internet Is Too Small

• Google to search data of millions who used ‘incognito’ mode.

• Italian Data Protection Authority latest to rule against Google Analytics.

Deals/M&A

• Propel raises $5.5m, a platform to build revenue-generating analytics products.

• Supersimple raises $2.m to rethink how companies work with data in the age of AI.

Data visualization of the week

• If you order Chipotle online, you are probably getting less food.

Smartest commentary

• “71% percent of respondents answered correctly on the number of pie charts they created in the past month (0)” -Jason Ganz, Senior Manager, DBT.

Datapoints of note

• Poor data quality emerging as a predominant issue for 57% of professionals, an increase from 41% in 2022.

That’s it,

-Ben

..

See our most recent on getting to page quality using attention metrics.

The Analytics Note Resources [continously updated] ← please do send any you recommend, books, podcasts, websites, even other newsletters.