It’s easy to get sucked in by vendors or evangelists who tell you that their tool, language, or platform will solve all your problems. But for most organizations, one size doesn’t fit all.

At the same time, you don’t want to let users and managers create a snarled rat’s nest of solutions that’s impossible to maintain or support. You need find a balance between chaos and order.

Given how fast analytics/data science tools and users’ expectations are changing, trying to plan it all out is a recipe for disaster. Instead of trying to create a single solution, incrementally, iteratively grow an ecosystem that will evolve and adapt to new circumstances.

Example: Tableau

Here’s an example of how a large organization might grow out one facet of their ecosystem.

You have a “on the trail, off the trail” policy:

  • Most users are “on the trail”: there’s a limited set of tools they can use, which are supported by your IT help desk
  • A handful of more sophisticated users are “off the trail”: they can experiment with new tools, but if they get into trouble with these tools and are stuck, they’re on their own – at that point, all IT will do is wipe their box and reinstall it using a master image (a practice the user and their manager have agreed to in writing)

Two of your “off the trail” users have been getting great results from Tableau, and they think your organization should adopt it. Working with your cross-department Tools Advisory Team, you run a short, time boxed pilot project and decide to invest in Tableau. From the pilot and ensuing conversations and analysis, you’ve figured out that you’ll need to support 2 types of Tableau users:

  • A small number of users are going to intensively use Tableau; many of them will also evangelize using Tableau and will provide a little mentoring/support in their department
  • Most of your users will use the same reports over and over. Some of them will rarely need to make changes, others will tweak their reports or create new reports every few weeks or months

So, you’re going to use a different strategy for each group.

  • Working with the Tools Advisory Team, you hold a one-week, on-site training for select group of users from several departments who are likely to intensively use Tableau. One month later, you follow up with a half-day workshop where the trainer remotes in to help users ungunk their biggest problems/obstacles. After that, you provide lightweight support, focusing most of your efforts on putting together good list of resources and nurturing a small community of these heavy-duty users, encouraging them to share examples, tips, etc. on your intranet and to help guide the future use, training, and support of Tableau across your organization as well as discussing what other data visualization tools/strategies the organization might adopt for addressing needs that Tableau doesn’t handle well.
  • A short while after you’ve trained your heavy-duty users, you begin rolling out Tableau to the rest of you users. These users get a short, 2 hour training. You also put together templates, cheat sheets, and other ways of providing on-demand support as their needs grow over time; most of the materials you use come from a cross-organization ecosytem. As part of your monthly power user brown bags, every 3-6 months you either offer tips or brief demos by other users of how you can get more out of Tableau.