This is the second part of our “Analytics team — Wins you can implement today” series. Part I about communication tips is available here.
All data analytics teams have different issues: how to set up a clean data pipeline, how to maintain it, how to share dashboards (bonus points when there will be actually used), how to limit repetitive questions from being asked, how to pipe back the data to operational tools, …
But in the end (and as we could expect), all teams want to maximize their impact without growing the headcount exponentially. Tooling and technical skills are a very important part of the solution but the obvious one. Less obvious and still pretty often the bottleneck, communication and knowledge make sure your team has the expected impact.
If you’re in an analytics team (or if you are the analytics team!) and want to boost your impact, there is every chance that you should start by tackling communication and knowledge first. What do I mean? Here are some tips you can implement starting today!
After browsing, crunching, and exploiting your data, you gain valuable expertise and knowledge about them, which should definitely be shared within your team and your company (for your future self, your peers, upcoming newcomers, your business users). How valuable is it? First, let’s have a look at what kind of expertise you develop.
Obviously, you get accustomed (some would even say intimate) to your data.
Here are some examples of apparently low-level, and yet critical!, elements you get overtime:
You also develop knowledge around the way to prepare, plan and run specific analyses, either based on their topic (user retention, acquisition campaign performance, etc.), or based on the statistical tools they involve (clustering, survival analysis, etc.). You know what steps to go through, what decisions to usually take, what assumptions to reasonably make.
Yes, you are good at data!
But this is not the end of the story.
As a data team member, your goal is to irrigate your company with relevant data so that your business users can make the right decisions. This means raising the bar (in terms of cost, quality, delay) for everyone. Leveraging every ounce of knowledge and expertise you develop is key in this perspective.
While this sounds like a quite large and intimidating topic, it does not have to be complex. You do not need to implement a company-wide exhaustive data cataloging solution, associated with numerous internal MOOCs and Ted Talks!
Here is a selection of wins you can take today.
Difficulty 2/5 — Costs 1/5 — Impact 3/5
You are particularly proud of your last analysis on retention? Think it implements many best practices your team should follow and could serve as a template for the next ones?
Make it available for everyone! Decide on a shared location, and start building your data A-list, to be read by any newcomer in your team.
For ideas about where you can post, a Slack channel, or a bi-weekly newsletter might be good to start. Check the first part about communication here for more details.
Difficulty 4/5 — Costs 3/5 — Impact 4/5
Data as a domain moves extremely fast. For you data analyst, it’s critical to stay at the edge of that field, but everyone else in the company cannot do that. Organize regular classes open to everyone (at least bi-annually). The goal of those classes is to show what the data team can do for them and help them adopt a data-oriented approach.
The difficulty here resides in the variety of profiles you may attract (but it’s crucial to welcome everyone). Carefully listen to your audience, understand what they are looking for, and help them grow.
With these classes, business users should better understand how to use and interpret data they already have access to, better express their needs when they come to you, and demystify the most recent trends — “No, we can’t answer everything by simply putting some AI on top of our database, but yes, last developments in natural language processing could be extremely useful to the marketing team, and we can work together to identify opportunities”.
If you run data classes, prefer subjects around use cases that make sense for your audience, over generic training session on tools.
Prefer subjects around use cases that make sense for your audience, over generic training session on tools; share your own experience:
No need to start big: sharing the learnings of a recent analysis might already be a good support for a first session. Also, announcing it publicly is a good way to engage you and your team in this virtuous circle.
Difficulty 1/5 — Costs 2/5 — Impact 3/5
If you are in a tech company, chances are that there is some software development team around you. Guess what? Knowledge management and sharing is also a topic on their side!
Software development and data teams have a lot in common (and also big differences, we won’t enter the details here!). And while I don’t believe there can be a one-size-fits-all way of handling knowledge (because you want that knowledge to be actionable, it needs to be bound to your own ways of operating), there are definitely best practices that can be applied to both worlds — think about sharing code snippets, or maintaining technical documentation for instance.
Don’t stop here! Do the same with more distant teams that do build and manipulate knowledge a lot (and sometimes for much longer in your company). Think about R&D, QA, Operations.
Get up from your chair and go talk to them (or give them a call if you are working remotely!)
In the information age, knowledge as a whole is a key asset for modern companies. It is even more important for a function and culture that is meant to penetrate each and every part of their activities: data. As a member of the data team, to fully achieve their potential, you have the responsibility to share and build upon your expertise.
Last but not least, it does not need to be a big company-level project, with even bigger change management processes. Short-term wins can be easily implemented, and already deliver value to everyone. The best way to build them durably is to seize every opportunity in your daily work, and iterate within short feedback loops.
Nothing new here, but worth remembering: start small, get feedback, iterate.