Every company today seems to be consumed with building an AI strategy. From sole traders to massive organisations, the question top of mind for the past 6 months has been – how can I use AI in my business? Often even before "Should I?"
Luckily, with an explosion of new tools built using AI it is much easier for an individual or company to incorporate AI into their day-to-day than ever before and it makes good sense for all of us to constantly consider whether we are doing our work in the most efficient way using the available technologies. But when it becomes a significant investment or derailing priority then a deeper reflection is needed.
I’m a strong believer in using technology, AI and Data to further your existing business strategy and priorities. To enable your competitive advantage rather than to be priorities in and of themselves.
When I think about AI and its potential for success I look at it in three parts – yes of course the machine learning and models that seem to get all of the buzz and attention, but equally (if not more important) the people building or using it and finally, the data and infrastructure it is built upon.
So if you’re a business leader considering what your AI strategy should be for 2024 then I’d encourage you to first consider, what is your current Data Strategy?
Do you have something at an overall company level that you can confidently call a Data Strategy?
Does it need attention before you should consider building a complex AI strategy on top of it?
The exercise of stepping back from an individual data or analytics problem, whether as narrow as a channel specific Marketing measurement question or as broad as how AI can transform how you do business, and instead taking a birds-eye view of how your company uses Data today and how that can be improved is something that I am always drawn to. I think it is one of the most worthwhile things a leader can do to consider what are the barriers TODAY to data being the silver bullet it was always lauded to be and to address those barriers before adding any new bells and whistles on top. I know this isn’t the "sexy" (I hate using that word in this context) or exciting project that makes shareholders wake up in a board meeting but it might be the work that enables the results they care about.
In this vein, a few years ago Tony Evans and I did just that, having worked on analytics and data projects in different teams and with various clients for a meaningful period we paused. We thought about the breadth of challenges we were seeing and mapped out a framework we called the 4 S’s of Data Strategy (the 4 S’s were admittedly for our entertainment as much as anything else).
Ultimately this is a framework that classifies data sophistication of an organisation under the headings of:
- Data Availability and Structure
- Analytics Strategy and Planning
- Data Value and Direction
- Careers and Talent Acquisition (People).
It merely scratches the surface of all of the various states that a company can be in under these headings but seeks to serve as a thought starter to consider where you are in each respect and whether a top-down approach to Data Strategy, that isn’t driven by one team or one division, could bring your organisation to a more streamlined and impactful place in deploying data for business results.
Data Availability and Structure
This section is the home of many a "digital transformation" project for the past 10 years. I really loved the metaphor Kelvin Gillen used for Data Lakes as I think it perfectly encapsulates the mindset and behaviour that surrounds a lot of data collection. While regulations like GDPR might have since put manners on this practice (in the EU at least) the "just in case" mentality hasn’t completely disappeared as more-is-more continues to be the prevailing attitude when it comes to data value.
The biggest takeaway here is that nobody has any business considering a complex or company-wide AI project if you haven’t done the foundational work on data infrastructure, data quality and data governance first. It would be like building a house of cards on quicksand.
Analytics Strategy and Planning
Who decides what analytics work gets done when and by who? Of course you want to empower your very talented Data Scientists to propose projects, but you can empower them more if they know how that work ladders up to the company mission and they are connected into the discussions and decisions that allow for that.
We have all had conversations about prioritisation between Product and Marketing functions (to name just one pair). It’s not on the roadmap. It won’t get prioritised above keeping the lights on or daily reporting. Sound familiar? Of course it does. Because Analytics planning is too often about who shouts the loudest or who owns the resources. I have yet to see a leader whose role is to look back after a quarter or even a year and assess whether the analytics work done was at a minimum aligned to the company priorities and even better (but I know not always easy) delivered measurable impact towards them.
Data Value and Direction
What role does Data really play in your organisation and how you make decisions? Are you paying lip service to its importance or have you really thought about its seat at the table? We are all susceptible to "gut feeling" and confirmation bias. It would be naive to pretend they will disappear completely, so we need to be intentional.
So much of this is about the real operational use of data in your organisation. Incentives and KPIs for individual teams. Communication of success metrics across functions. How learnings scale when we have them. Our threshold for what evidence is enough before a change is made. Having common languages and common processes even where nuance exists means data can really embed as a valuable input.
Careers and Talent Acquisition
People are everything. If you want good people you need to create an environment they want to be and stay in, one in which they can thrive and do their best work.
Too often I see companies rushing to hire Data Scientists who want to work on the most sophisticated models and then get frustrated either because the infrastructure doesn’t exist for them to do what they do well or the modelling projects are for their own sake and so don’t end up being deployed or having the impact that they could because their output isn’t a priority. Yes Data Scientists want to hone their technical skills but they want their work to make a difference, hire them when they don’t have to choose. Vanity hires, in most cases, do not result in long-term satisfaction or mutual success.
I have no doubt that there are companies in the Starting or Surviving buckets that have highly skilled and well run analytics teams, but that is very different from having seamless integration between data/analytics and the business objectives where everyone across the company can tap into that equally. You might be lucky to have a superstar of a data scientist or analyst who does a lot of this work themselves but what happens if they leave? Do you have the processes and culture in place for these benefits not to leave with them? Do you have a structure that will allow for ongoing progression for those that are ambitious? Do you have the right problems for them to work on so that they actively want to stay?
Think of needs in a hierarchy
You’ve probably heard of Maslow’s Hierarchy of Needs. A few years ago I came across a Data Science equivalent here. I think it’s a great framework not just to consider what your company needs to be doing as data fundamentals but equally to consider where you need to hire and when, starting at the bottom rather than skipping to the trends at the top.
If you want to use a self-serve AI tool for productivity today, by all means do that. But don’t hire AI engineers (unless your core product requires them) if you can’t yet deploy insights effectively across teams.
So what?
As with any framework, neither the 4 S’s or the Hierarchy of needs are perfect or all encompassing. But hopefully either, or both together, can be a useful tool to kick off a diagnostic exercise in your firm.
What are the dysfunctions that exist in your data or in your teams today?
If you have big dreams for what you can do with data and technology, why aren’t they being realised? Or, likely, why are you faltering inmuch smaller projects?
Artificial Intelligence, Data Science, Data usage in general are more than a hype cycle. They can be genuine enablers of your strategy but if you’re struggling to implement simple capabilities then you will equally (and more expensively) fail at more complicated ventures.
Teams and individuals can’t reach their full potential as islands. As with any significant change you want to make, buy-in and communication are paramount and a shared understanding is a great place to start with a data strategy. Where do you want to get to? How can you improve the connection between teams, truths and technologies?
None of the component parts categorised above happen by accident but mapping them all out with a warts-and-all view of where you currently are can help you move each one forward, future-proofing your business for the next big thing while you do.
Uses some of the work and thinking carried out with Tony Evans while working in Meta.
Having a robust Data strategy that delivers business value can be a multiplier for company performance.
If this is something you or your leadership team need help with then check out my offering on kate-minogue.com
Through a unique combined focus on People, Strategy and Data I am available for a range of consulting and advisory engagements to support and enhance how you deliver on your strategy across Business, Data and Execution challenges and opportunities. Follow me here or on Linkedin to learn more.
