Create Data Strategies That Actually Delivers Business Outcomes
6 Practical Tips for Chief Data Officers
I once hired an accomplished data strategy leader, let’s call him Pete (from Pete the cat of course), to develop a data strategy for us. The ex-big4-consultant hit the ground running, interviewed scores of internal stakeholders at all levels and quickly developed a draft strategy for my review.
The document had more than 200 pages, with more technical details than required to run a nuclear plant. Pete had included a roadmap but all I saw was a bunch of tech projects for the first 18 months, and no articulation of what business benefits will be delivered, how and when.
Pete was quite proud of his creation, but what I saw was a strategy created around a wish-list of things and lacking a clear vision. The things in the wish-list were low value, non-strategic, and things that would have gotten done despite a strategy.
Pete had listened to stakeholders, created a technically perfect blueprint and delivered a very comprehensive document for sure, but in doing so, had missed an opportunity to drive innovation and strategic impact on the business.
Pete had fallen into the classic ‘give me a faster horse’ trap!
Signal from Noise
Your data & AI strategy shouldn't be about data, AI or technology at all — it should be about specific business objectives & driving your business strategy
Your strategy should be simple and understandable by all - stay away from buzz words and complex articulations
Culture eats strategy for breakfast - make sure you address cultural and literacy barriers
Behind the Dashboard
Unfortunately, Pete's story isn’t unique in the industry.
His strategy document was packed with technical jargon and complex data architectures but didn’t clearly articulate business outcomes his strategy would deliver. While there was a roadmap, it was a huge waterfall style plan which would take 5 years to implement, filled with mostly technology initiatives, and no clear path to delivering early and incremental business value.
Even with my technical background as an engineer and an architect, I was finding it hard to make sense of the content and flow of the document, let alone understanding what value I would get by spending a cool mid-eight figures.
Worst of all, there was no reference to the change management, uplifting data culture and uplifting skills of users - the strategy simply ignored the significant organisational change required for the strategy to be successful.
So how can you avoid Pete's pitfalls and create a truly impactful data strategy? Here are 6 practical tips, all from my real-life experiences, that will help you navigate common obstacles and develop a data strategy that drives real business transformation.
Playbook || Data -> Insights -> Action
1. Your strategy should contain mostly business stuff, almost zero tech stuff
I remember sitting through many data strategy presentations (and having delivered a fair few myself) where 45 minutes of the presentation was spent in explaining the difference between a data lake and a data warehouse.
Nobody cared. And it didn’t matter.
What matters is how data and analytics can help improve accuracy of liquidity forecasts so cash can be better invested, or how can data analytics reduce churn in your high-income customer segment, or how industry or technology disruptions can be leveraged to improve customer service, increase revenue or create new offerings.
Majority of your strategy needs to focus on big business problems, and how data analytics can help solve these problems. While technology is important, it’s just an enabler.
This sounds pretty obvious, no? Sadly, it is not. It’s amazing how data strategies don’t even mention any business outcomes they are intending to drive using data analytics.
I’ve learnt from painful experiences that Executives don’t fund fancy tech stuff - they fund delivery of business outcomes. If your data strategy is 70% tech content, you have a s**t strategy.
These days I tend to include only 1 slide on technology and architecture in a data strategy pack. And this slide is almost childish—coloured boxes with simple arrows, stripped of all technical sophistication. My former architect self would be horrified, but this simplicity gets just the right amount of detail across to the people signing the check.
2. Keep it simple, stupid!
The most effective strategy I wrote was all of 3 slides long, and it wasn’t even called a strategy.
It didn’t have a single reference to a tech trend, had no unnecessary visuals/diagrams, no case for change and almost none of the obligatory strategy slides.
In contrast, my worst strategy was technically perfect 152 pages long (although we had a summarised 30 pager) - 30 pages on the context and case for change, 20 pages on architecture and another 23 on tech tooling. A big 4 partner would have been proud of this document and would have charged $1m for this, but this was too much, too complex and too technical for it to be effective.
Keep your strategy so dead-simple even your 15 year old will understand.
3. Be Strategic - This is a f***ing strategy after all!
Strategies should strike the right balance between ambition and realism. While there are risks in a strategy being over ambitious - unrealistic goals, lack of buy-in, risk of burning people out etc., to me an unwillingness to be bold, innovate and push boundaries poses much greater risks.
Think what Kodak’s reluctance to embrace digital photography did to them.
A strategy that only focuses on improving operations (See Zone 1 RISE Framework), excludes impactful initiatives because of budget constraints, and avoiding disruptions like AI is a ship anchored in a dry dock while competitors sail into the horizon—secure in the short term, but destined to rust as the tides of innovation rise.
Inject some ambition into your data strategy and set bold ambitions. If data analytics can really help your company make or save more money and you can articulate how, securing funding is possible. If there are cash constraints, think about slowing or staggering execution - not everything needs to be done in year 1.
Don’t restrain your strategy based on your current constraints, optimise for future needs and position data analytics as a strategic driver.
4. Balance internal voices with external insights
Pete over-indexed on feedback from internal stakeholders, many of whom had been at the company for decades (some for as long as 45 years!). Many who provided feedback had no knowledge or experience delivering or being part of data transformations in other companies. They certainly did not know what good looked like.
Company insiders, especially business leaders & long serving employees often can't see the forest for the trees. They tend to focus on current pain points and short-term constraints.
Am not saying you need to ignore internal stakeholder feedback, but this is only one side of the coin.
You should make a concerted effort to incorporate external viewpoints. Talk to your vendors & consultants - they are often miles ahead of your company, and they work with lots of companies and can tell you what works and what doesn’t. Talk to companies in your or other industries - learn from other’s successes and mistakes. Engage with academia and industry groups - they study success stories and often have frameworks you can use to drive a successful strategy.
5. Include a roadmap, but not a monolith
If I had a dollar for each time I heard ‘but I can’t deliver any business outcomes until we build out full data foundations!’, I’d be a gazzillionaire.
Yes, building stuff takes time, especially when you are starting from a low base. But it’s not a great situation where data foundations take 18 months to build and only then will you start building business use cases. No one has the patience to wait for 18 months while you spend millions of dollars.
When designing your roadmap, make sure you find a balance between building foundations and delivering business value. Break work down into smaller chunks and deliver slivers of value along the way. Deliver small units of features and iterate on this towards a full solution. Incorporate wire framing, prototyping and feature demo’s in your delivery lifecycle. Stay close to key business users and release features to them early and make sure you incorporate feedback into the next delivery cycle, creating a virtuous cycle. Instead of building your roadmap once upfront, continually build and update based on real-time user feedback.
I know some people struggle with the idea of not having a 4000-line Microsoft project plan upfront that describes all work packages, activities, dates, milestones and dependencies.
But building adaptable roadmaps that you can iterate on continually will give you the ability to improve time to market, reduce implementation risk and gain executive confidence as they will no longer view your project as a black box.
6. Address cultural and people challenges
For the longest time I ignored the cultural and people aspect of data. I pushed fancy strategies, and brute forced my way through several transformations. I viewed change management as a ‘give users a warm hug and make them feel better’ and frankly unnecessary thing - why did I need to treat my users this way, they were professionals after all.
Ah those were simpler days, when I was young and foolish, and I could blissfully ignore things that I deemed unimportant.
Obsessing over delivering a technically perfect solution is fine, but eventually it is humans who use your solutions. You are not just implementing technology, you are asking people to fundamentally change how they work (make decisions, operate processes etc.). Your transformation project will challenge and change people’s existing identity, autonomy and status (real or perceived).
This is especially true in today’s age of AI where many roles are becoming redundant, ways of working are evolving rapidly and the entire shape of the workforce is changing.
People will feel threatened, and misaligned incentives will need to be identified and changed. If you don’t have your users on your side, you will fail. Hire a competent change manager (usually the first role I hire). And get into the habit of dishing out “warm hugs”.
Your strategy needs to open with culture and people, and address what it takes to align culture and people to your strategy. Technical debt can be refactored or risk accepted. Cultural debt will compound exponentially and ultimately lead to the collapse of your data transformation.
The final word
Look, I’ve been there. I have created data strategies that resembled technical manuals. These six tips I've shared aren't theoretical - they're battle scars from years in the frontlines.
I’ve gone from a tech-obsessed geek to a business-focussed strategist who has learnt (the hard way) that data strategies need to address core business problems, simplicity trumps complexity, and your success almost entirely depends on how well you handle cultural & people change.
Don't be like Pete. Be better. I certainly had to.