A Manifesto for Radical Analytics
Stéphane Hamel
11011

André Mafei: thanks for being the first to chime in! You see, this is a point where I have a different opinion, and why I left out any technical/tools aspect from the initial premises of my manifesto. In my experience, and based on the digital analytics maturity work, “technology” has very little to do with success (ok, ok… up to a certain point!). The concept of centralizing data is a recurring topic since the 1960s and the idea of “data-warehousing” has been à la mode since the 90s. Now it’s just more trendy to call it a “data lake”… My first “central database” project goes back to 1987… the concept haven’t really changed. Big Data didn’t give birth to the idea — it just made it more accessible and easier (in theory) and allowed the so called “data scientist” to claim a new buzzword. The truth is… the concept of a central location for all data is a fallacy! The definition and concept of Big Data itself stipulate data will continue to grow exponentially, come from new, unpredictable sources, and morph depending on the needs at hand. To use the analogy, the data lake will never be filled. Merely replicating data for the sake of bringing it into a data lake doesn’t add any value — to the contrary, it creates a ton of additional issues (ex. why bring GA data into another infrastructure… or why would you take it from Big Query and put it anywhere else?). Instead, the analyst needs to be able to connect to the right data at the right time — something tools like Tableau or even the new Google Data Studio and other tools are trying to do.

Guillaume Martin: “difference between reporting and analysis” and “actionable report and deep dive analysis”. We heard those many times before, and our brain accepts the idea as being very logical and the way to go. The problem is we collectively lack the “how to” — we have those nice statements that makes perfect sense, but when you are back at the office, how do you actually create the conditions to become a reference and a leader, instead of digging your hole and becoming a reporting monkey, sorry, I mean provider… Thus the premise of “Never ask, always propose”, and other techniques, like active listening. The thing is… there are plenty of resources to tell you how to use GA, but few resources to tell you the soft skills require for the job.

Alex Mylopoulos: online-offline, multichannel, cross-device, and the adtech and martech that are a real mess with thousands of solutions. You probably have seen this crazy landscape supergraphic

At one point in my career I was asked to list all the programming languages I used at one point or another. Assembler (yes, I’m old!), Basic, Pascal, Cobol, Fortran, C, C++, Lisp, PHP, C#, Java, JavaScript, Perl, Arc, Bash, C-Shell, Clipper, DOS Batch and certainly others I forget — and that’s not counting the databases and tools! Gee! I even worked in the Computer Aided Software Engineering where I created a software in C language connected to an Oracle database on Unix that generated complete applications in Cobol, for DB2 database on IBM mainframe (because at the time, developing on mainframes was too costly).

Why am I saying all of this? Because once you master the concepts, learning a new tool is easy!

Mohamed Hédi Lassoued: at last! There’s light at the end of the tunnel. But seriously, how many analysts learned about Agile, Kanbans, customer lifetime value and basic marketing foundational concepts?

Anonymous #1: a person who will remain nameless because he works for a vendor in the space, wrote “I feel like I’m the only sane person in the nut house (or vice versa) when I discuss these things with fellow analysts, or management, or just about anybody. I just get blank stares or quizzical looks, which tell me that people are not thinking about this stuff at all. … just license a tool(s) and hire somebody who knows the tool, and bingo — you have digital analytics”. This person have a strong background in CRM/database marketing and feel he has “been exposed to up-sell, cross-sell, churn/loyalty, propensity, and RFM models, which the vast majority of digital analysts haven’t, since their only exposure to customer or prospect data has been through IP addresses and cookies, which is quite different.”. In conclusion, he agrees the distinction of “digital analytics” will fade away.

Keep the comments and feedback coming!

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.