BD #4 - The 4 Question Data Strategy Framework
A proven system for building an actionable data strategy quickly, in any context
Prompt: stepping stones to clarity (made with Midjourney)
Name change, the polls have spoken - A Week in Data is now Beyond Data.
Also, announcing our new referral program where you can earn free gifts. Login to the site for more details.
In today's issue, I'm going to share the framework I've used to develop data strategies for tens of companies - from small startups and scaleups to national organisations.
Ask any seasoned data professional what to do when starting on your data and analytics journey and you'll hear grumbled cries of "you best have a data strategy".
The thing is if you've never built one before then where do you start?
Google will point you to big consultancies and experts willing to sell you their expertise. Now don't get me wrong - I believe this could actually be some of the best money you'll spend. Building strategies that are actionable is difficult. The pace of change and the blurriness of terms in the data industry makes this even harder.
But what if you can't afford to bring in these experts - what if you need to prove some value to get the bigger business case or convince the budget holder?
Well, I'll walk you through what I've done many times before (and the core of what we do at Hypercube) to arrive at a data strategy that gets things moving in the right direction.
The pillars of a great data strategy
This approach hinges on four questions that will guide you through gathering the right information and pointing your data and analytics efforts in the right direction.
A note about this approach - I've had the most success with organisations < 250 people all within the same country. Above that number things start to strain and get difficult to get the right people in place. If you're in a much larger place or across multiple timezones it might be best to do this for sensible sub-sections of your organisation.
This approach is meant to be agile, lightweight, and quick - prioritising action and iteration over big up-front planning. The goal is to get enough of an understanding of where to focus your efforts so that you can prove out the value and drive results - refining this strategy over time to stay on track.
Step 1: Where are we now?
The first step is all about gathering the current state of play. You want to do a lot of information gathering and documenting your findings upfront. There's nothing worse than starting some big band initiative that some forgotten stakeholder on the other side of the company has already delivered.
You absolutely want to align with the wider business strategy and things like the tech and commercial strategies (if they exist? I hope they exist).
Think here about the dimensions of what really matters to your data strategy. There are a lot of ways to cut this but what's important will depend on your specific context.
As a minimum I would consider:
Security and compliance
Governance and observability
Architecture and infrastructure
Storage and operations
Analytics and reporting
Sources and sinks
People and skills
For each of these, define what matters most to your organisation and then document the current state of play for each - who are the stakeholders, what's been tried before, and why are things the way they are.
The best way to get this information is through forms sent via email with workshops to follow up and expand on any points of interest. Involving as many people as possible encourages ownership and improves awareness of what matters from a data perspective.
Step 2: Where do we want to get to?
This is usually the fun bit.
Find the highest value and highest risk parts of the business and start there. Get stakeholders to share their thoughts and insights on what they see are the biggest opportunities for improvement. The more the merrier. We'll prioritise later.
Capture ideas for projects, tools, and process improvements and group similar suggestions. Take note of who suggested them and what part of the organisation they should impact.
I also like to do "Art of the Possible" sessions here - take senior stakeholders into a workshop and present to them what the best in class, state of the art in their market are doing. Bonus points if you can show them where competitors have an edge they aren't exploiting.
Step 3: How are we going to get there?
You now understand the business in its current state and what many of the people across the organisation think should be done to improve things. You're now going to want to come up with rough estimates for how much effort each of these initiatives will take.
Firstly, categorise each of these ideas into the following:
Core - things that deeply differentiate this business, fundamental to its existence
Complex - things that are difficult for this business and its competition to deliver
Commodity - no edge to be gained, easy to acquire and maintain, non-differentiating
For most businesses, you definitely have a strong case for doing anything Core in-house. This might mean spinning up whole new functions, but the point is delivering this kind of capability should significantly drive the business towards achieving its wider ambitions.
The Complex initiatives are interesting and decisions should be made on a case-by-case basis. Sometimes these things can become Core after a few years once they've been delivered successfully. Be careful here though - this is where pet projects and wasted effort lurk. If you don't already have the capability in-house then I would stay cautious. Reach out to external parties and experts first to prove value in these and then bring the skills into the business. This might seem expensive upfront but it pays off compared to hiring new teams for projects that never get off the ground.
With the Commodity projects, you want to find a good tool, a third-party managed service partner, or a small training budget to get these done as cheaply as possible.
Once this categorisation is complete - work through each initiative with the relevant delivery model and work through estimates for how much each will cost and how long they will take. Make note of any dependencies between them and also where low-hanging fruit might appear when one is done prior to another.
Step 4: Where should we start?
Now you want to make a plan. Take each initiative you've outlined in the previous steps and have senior stakeholders rank them by potential impact to the business strategy. Ensure they're considering value from a monetary and cultural perspective, as well as any impact on risk factors outlined earlier. You want to keep the group much smaller at this point - key decision makers that are going to own and champion this data strategy with you.
Pay particular attention to any potential synergies between them.
With that ranking, you can prioritise an order in which to deliver things. Use the age-old 2x2 grid to plot them by impact against the effort to implement.
Deliver, refine, iterate, succeed
Now if big waterfall projects and Gantt charts are your thing, go ahead and figure out where and how these will be delivered. I
I prefer doing something smaller and delivering incremental improvements - because often, priorities and context change much faster than our plans can handle.
Take the biggest impact, lowest effort piece and build from proof of concept > minimum viable solution > business as usual solution in short, sharp delivering iterations. At frequent, regular intervals, reprioritise each initiative against the rest and see if more improvements to the current workstream seek to add more value than building something else from the list.
If you can do multiple things in parallel or supplement some of the easier pieces with training existing teams - all the better.
Keep doing this and reevaluating what's on the list and over time - you'll be building data solutions that significantly impact your business with limited risk of going off course.
I hope that's helpful. It's worked for me plenty of times and it should be enough to get started on a data strategy that's right for you.
When you're ready, there are a few ways I can help you or your organisation: