How To Stop Your Data Team Becoming A Bottleneck
When it comes to incorporating data/ insight into your business, most companies have a pretty simple formula in mind:
A centralised data analyst/insight team becomes the go-to source for information in your organisation. They help to flesh out ideas by looking into the current data, fill-in gaps through sourcing of third party data/market research. They help you make the right choice. It seems pretty simple and yet the reality is that this model can actually cause a lot of problems…
Analytics teams typically have a small headcount and multiple stakeholders. What happens is that those multiple stakeholders all have needs at the same time, and so the insight team quickly becomes overloaded with work and will have to juggle multiple priorities. Not only does this mean that sometimes the accuracy of their work is compromised but more often, they are so busy putting together reports, they don’t have the time to construct meaningful suggestions/observations/stories.
Rather than helping your company become more agile, they actually become a blocker in the whole process, slowing it down. This is especially true where their jobs may be reliant on an understanding of code, programming language, or niche software which means other people cannot easily access the information.
It’s also a problem which builds over time. The more data initiatives you undertake, the more dashboards, models, tables and infrastructure which all need to be maintained and the less time your team spends providing value.
“My controversial opinion is that some people like to be in this position.” says Oliver Gwynne. “If everything needs to go through them, it makes them feel like they are indispensable to the decision-making process, but this may not actually be the best thing for the company.”
How To Tackle This?
In order to avoid this issue, you need to put insights and data directly into the hands of decision makers, in tools that they actually use. You have to make it as easy as possible for people to access relevant information and you need to train them so that they can read, interpret and act upon that data.
Data should not just be the remit of the analytics team. Self-serve dashboards are a great place to start but data models should help to facilitate innovation beyond reporting. Think about automated lists, flags, push notifications etc. Any way in which your analytics team can help make decisions easier without people needing to come to them for analysis.
We have often found that people often think that with “bigger” decisions (or with decisions which maybe have bigger budgets associated with them) that the more information they can get, the better. But more information often clouds decision making and leads to ‘analysis paralysis’ whereby you can’t see the woods for the trees. It is a data leader’s role to help stakeholders think about which core pieces of information they really need to inform a decision. How can they do more with less? Otherwise data teams may spend a lot of time gathering 65 points of information which all confirm the same thing.
While it is natural for growing companies to first adopt a small, centralised team, as your needs grow we would recommend employing a ‘hub and spoke’ model whereby half of your analysts’ time are spent with different business verticals to help democratisation efforts.
Data Champions, individuals who are passionate about using data effectively and who promote data best-practice can also be very useful. They don’t need to be technical people, just people who can ensure people are using data in the right way, thinking carefully about their requests etc. They can also act as a ‘first line of defence’ in safeguarding your analytics teams’ time.
Conclusion
Any department with a small headcount and multiple stakeholders is likely to become a blocker in your processes. For analytics team this may mean that they slow down key decisions, rather than making you more agile, and the more tables, dashboards etc they create the larger the technical debt and time they will need to spend more time maintaining them.
Their first priority should be automated reporting and self-serve dashboards and from there, consider data champions and a hub-and-spoke structure so that data is embedded in your business functions. The best way to avoid ad-hoc requests is a formal data roadmap tied into business strategy.
If you need help understanding business priorities, creating a roadmap and supporting your analytics team…contact 173tech today!