One piece of that broader strategic approach involves understanding how organizations evolve over time and using that evolution to shape immediate tactics and investment decisions. Models of evolution are widely called “maturity” models and they are ubiquitous in both strategy and sales discussions. Senior Leadership wants to know “where they stand” and which investments will truly drive the organization forward.
There are probably as many Maturity Models as there are consultants; indeed, it's quite possible that like Internet-enbabled devices, there are more than one per person.
The one I tend to use was adapted from a model that I believe was first created by Stephane Hamel many years ago (I’m sure Stephane has changed it since then so don’t blame him for anything I say here).
The basic idea behind the original maturity was simple. Organizations started out focused on Site Analytics (Stage 1) and gradually evolved into more complex but more impactful analytics involving increasing levels of segmentation and integration (up to Stage 5).
I think this is still a fundamentally sound model though there are far more branches than this simple story would suggest. We’ve grown a lot in the past half-decade. There are so many differenttypes of analytics and so many different approaches to things like big data that any fairly straight line approximation of maturity is necessarily going to simplify things greatly. What’s more, as every organization I talk to further reinforces, the typical enterprise is not at one place on the curve but at many different places in different parts of the organization or with respect to somewhat different problem spaces. Most organizations, for example, are at completely different points on the curve when it comes to Mobile App measurement or Social Media Measurement compared to traditional Web measurement.
When I draw that maturity curve these days, however, the two biggest changes I tend to make are to the shape of the curve and the stages that exist prior to site analytics. We still see a significant group of enterprise companies that are struggling with foundational activities around data collection and governance. Well – almost everyone struggles with these things – I mean enterprises that are struggling to the degree that they cannot expect or drive value from any higher-level activities.
Even more prevalent – maybe even in a flat majority – are organizations that capture some business value from reporting but that do essentially no analytics (site or otherwise). Because that’s such an important step in the maturity curve, I think it’s critical to represent it.
In this rendering, there are three stages prior to any real analytics but that are foundational and necessary. You simply can’t ask interesting analytics questions until the organization has achieved a level of comfort around digital data that only comes from widespread usage of customized reporting. And that widespread usage of reporting only comes from a solid foundation of fairly customized data collection.
The other distinct feature of the maturity curve drawn in this fashion is where I’ve placed the gaps. I draw the curve with a large gap between reporting and analytics and another big space between site optimization and Segment Targeting. I draw the curve this way because my experience suggests that these are the places where most organizations really struggle. Going from reporting to analysis is the single biggest jump an organization can make in digital analytics. Going from analytics to targeting is the second hardest jump. Both are critical to achieving analytics that actually matter.
Here at EY in the broader Enterprise Intelligence Group we sometimes talk about analytics maturity in terms of a shift in analytics complexity from BI to Predictive and then to Prescriptive analytics:
This is a different sort of approach – one that’s meant to
capture a broader range of analytics problems and is rather less specific to the natural evolution of digital analytics. Again, I think it captures a
pretty fundamental evolution (though – and this true with the other approach as
well – you never stop living at multiple points on the curve; BI doesn’t go
away it just becomes a smaller piece of the entire program).
I also find it congenial because it captures perfectly my latest thinking on reporting in the digital realm and the evolution of reporting from “historical” to predictive and prescriptive.
Our recent projects aren’t so much reports as business tools with embedded models that allow for forecasting, prediction and prescription in one tightly integrated package. On the other hand, this maturity model is missing any description of the evolution from foundational collection and governance to usage and it doesn’t embody the journey toward increasing levels of segmentation and personalization (quite naturally, since that journey doesn’t apply to every type of analytics).
Which brings me to an interesting discussion during and right after (you know a conversation is good when you just can’t let it end!) an X Change Huddle with Colin Coleman that focused on how organizations are evolving in the data warehousing space.
Colin drew a picture like this:
In Colin’s diagram, organizations move from siloed reporting tools toward integrated, in-house tools for reporting. From there, they then move toward increasing analytics – likely with a siloed big data analytics project. Finally, they shift toward a fully integrated big data stack (that starts at the detailed level of data and builds up) across multiple siloes.
As with any maturity model, not every enterprise is going to follow exactly this path and the diagram is somewhat specific to media. But I think it’s a powerful way to understand how most organizations evolve analytically and what seems like the most natural evolution from siloed reporting to integrated reporting and then to analytics.
Here’s another view of Colin’s model that focused more on the types of projects delivered within each quadrant:
I like this view very much since it provides a natural tie between the level of systems maturity and the level of analytics maturity you’re likely to see. Put the same curve over it that we used on the first chart and you have a pretty good trajectory for the vast majority of serious enterprise analytics programs.
So why does all this matter?
Drawing these maturity curves isn’t post hoc theorizing. Just as you can find the easiest slope on a hill by following the worn path of deer footprints, understanding the common trajectory of organizations can help you understand where the easiest path to evolution lies. This natural path can server as a kind of trailmap for your own strategic planning. If you’re trying to decide between Integrated Reporting and Siloed Analytics, for example, Colin’s model might incline you toward the former as the most natural path.
That’s beneficial, but it’s not the most important use of this type of thinking. What I really like about this kind of maturity mapping is that it helps guide deeper thinking on what types of work will provide the right foundation for moving up the maturity curve and creating the most natural and efficient trajectory.
For example, I believe strongly that incorporating segmentation into EVERY stage of this journey is appropriate. Colin’s model has segmentation in the most advanced, upper-right quadrant. As an historical description, I don’t disagree at all. That’s exactly the way things usually play out. But if you take heed of where you want to get to, it’s perfectly possible to build segmentation into your siloed reporting way down in the lower-left quadrant where most organizations start their journey.
The segmentation you build in that quadrant won’t be as mature or robust as when you get further along the maturity curve, but if you START by embedding segmentation, your organization will move much faster along the curve and be far more ready for the next stage in the journey.
Maturity curves, in my opinion, are best understood in the context of where organizations tend to stop. These stop points tend to be places where there are strong discontinuities between the type of thinking in one stage and the next.
If you look at the work we’ve been pushing hardest on in the last couple of years (segmentation and embedded modeling), it’s work that is explicitly designed to embed concepts from the upper-right quadrant into lower stages of the maturity curve.
By embedding those higher maturity concepts into the earlier phases, we create deeper levels of integration between each stage and facilitate the transitions between them.
It was a great conversation – one that got me thinking and deepened my own understanding of enterprise analytics evolution and how we can facilitate our client’s progress. Just the sort of thing that earned X Change that 100% NPS score that I can't help mentioning again!
Comments