Executing a Data Strategy with OKRs
Writing a data strategy is one thing, making it happen is another. Here’s how Objectives and Key Results can help.
OKRs are the company wheelhouse
Objectives and Key Results (commonly known by their initials as OKRs) are a framework for turning strategic intent into measurable outcomes for an organisation.
Invented by Andy Grove at Intel and having their roots in the management science that preceded him, they have been around for a while and used effectively at companies like Google, Intuit, and MyFitnessPal. I’ve used OKRs at the last three companies that I’ve worked at. If you search you will find that there are acres of articles written about OKRs. If you want to go deeper then take a look at whatmatters.com and maybe read John Doerrs’ book.
This article won’t make much sense unless you know the basics, so here’s a potted overview of OKRs.
Objectives are written as a set of goals that the company or department wants to achieve over a given time horizon, usually a year.
OKRs are built throughout the organisation. They may start with the executive outlining some higher level strategic objectives which are interpreted and adopted by departments and teams across the company. It’s important that OKRs are not cascaded down mechanically. Instead, because OKRs are shared openly it is possible to build them bottom-up or even “middle-out”. This avoids an inflexible over-alignment of initiatives and the corresponding effort to track progress across multiple layers of interlinked OKRs. It also allows for teams to introduce local innovation whilst still remaining aligned to the overall company mission.
Taking an example from the world of marketing; a CEO will set goals for acquiring customers and the Chief Marketing Officer in turn will develop the objectives of marketing campaign reach and customer acquisitions that are expressed as the Key Results (outcomes) that will show that these goals have been achieved. These OKRs will guide the Marketing department to define a set of measures to show progress against the OKRs. Once the measures are defined then a set of marketing activities can be developed and planned within the marketing teams. The teams can then get to work, aligned to the greater company goals.
As stated above, this is not to say that OKRs are top-down dictated. There needs to be just enough alignment and traceability between related OKRs to ensure that the company is pulling in the right direction. Each team, department, division, or whatever organisational unit is in play, are responsible for defining their objectives, how they will know when they’ve reached them and what they will need to do to get there.
As OKRs are transparent across the company they can help plug the strategic planning gap. The Data division can see what Marketing are trying to achieve and what their intended outcomes will be. The Marketing Data team can plan and see what activities they can help out with. Importantly the Marketing Data team can add in improvement objectives that might make them a better team and provide a better service.
When they are done well, OKRs can focus a team on only expending effort on the tasks that really matter to enable travel in the desired direction of the company.
The focus of using an OKR approach is execution. They should be about what really needs to be accomplished. This is why they are great when partnered with a solid data strategy.
Company strategy is the map and compass
Successful companies execute their strategy. Very successful companies do this ruthlessly.
The last decade has seen a rise in the significance of data to enable the overall company strategy. It has become common to see companies develop a data strategy as part and alongside their broader business strategy. IN general this has been accompanied by the appointment of a new C-suite role, the Chief Data Officer.
Like OKRs, there are plenty of good resources on why a company should build a strategy for data. The focus of most data strategies is to transform a company to become data-driven. Meaning, that data is recognised as a valuable asset that can be turned into information to program the direction of the business.
This is powerful stuff. So how is it that many companies fail to execute on their data strategies ?
You don’t have to look far to find evidence telling us that the majority of companies are failing to become data-driven. TechCrunch [1] cites a Harvard Business Review article by Randy Bean and Thomas H. Davenport of a NewVantage Partners survey that found that 69 percent of the companies surveyed reported they had failed to create a data-driven organisation [2], despite significant investments in data technology and AI initiatives . The news gets worse as Bean and Davenport’s article cites a downwards trend from 2017 with more companies considering themselves increasingly driven by opinion rather than data.
As is often the case, there exists an execution gap between intent, as expressed in the strategy and action, as evidenced on the factory floor. OKRs can plug that gap.
We’ve started to do this at the company I am currently working at. The previous iteration of the data strategy was outlined in 4 broad themes. Appropriate metrics (as KPIs) were associated with the themes. The intent was laudable and was expressed in Performance Goals but the approach lacked a formal framework for execution. This year we are using OKRs to provide that framework.
A comprehensive data strategy should comprise of the following elements:
Alignment with Customer needs:
Assuming that your overall strategy to build products actually corresponds to what the customer actually wants (or you’ll be going broke pretty quickly!), then the data strategy needs to address the OKRs of the functions building those products, i.e.; the data team needs to provide support for the product, marketing, customer services, mobile, website engineering, etc. teams to meet their OKRs.
Data Platform Technology and Architecture:
A plan to build a robust platform of data storage, data feeds, visualisation and modelling tools and a supporting connective infrastructure.
Analytics
An ability to apply models and perform deep analysis on the data that you have.
Democratisation of data:
Making data available where necessary, cataloguing it, making it discoverable and well understood to encourage staff in the company to make effective use of it.
People:
Hiring and retaining top talent, developing the staff you already have, fostering a culture of technical excellence and collaboration.
Compliance/Governance:
Remaining compliant with regulatory data requirements and company policies with respect to data collection and usage. Having efficient and transparent processes in place to ensure data teams are applying regulations and policies when developing solutions.
Data Quality and Management:
Setting the standards and mechanisms for data to be trusted as it flows through the company.
Security:
Keeping in lock step with the enterprise’s broader approach to keeping data and systems safe.
Data Literacy and Culture:
Plugging the outputs of models and analytics into the decision fabric of the company. How to take data outcomes and operationalise them, turning them into actions for the business. The promotion of data as a first class concern for the company.
Expressing the data strategy as OKRs
Once these data strategy elements have been outlined then an OKR can be built to deliver each one. There’s not necessarily a one-to-one mapping of strategy elements to OKRs though. More often an OKR might cover more than one dimension of the strategy. The exact OKRs will vary according to the particular circumstances of your firm. Here are some simple examples to demonstrate the technique.
Company Strategy: Alignment with Customer needs
O: Acquire more mobile app customers
KR: 75% uplift in mobile app downloads
Data Strategy: Align with functional teams to meet company objectives
Data OKR: all customer journeys that lead to app store purchases must have metrics collection and analytics to measure progress or drop-out.
Company Strategy: People
O: Commit to developing our staff to reduce attrition and skills leaking out
KR: Staff churn is kept below 10% each quarter
Data Strategy: Keep our teams technically skilled, engaged and current
Data OKR: 75% of staff in our data teams successfully complete 3 online technical courses in a year
Company Strategy: Make more informed decisions on product feature development
O: Data plays a key part of the input to product development
KR: Use Lifetime Value (LTV) calculations as an input to the product owners who are developing product features to engage higher value customers
Data Strategy: Improve Data Literacy for Decision Making
Data OKR: The output of LTV calculations are linked to >200 feature development story points in the product team scrums
… and rinse and repeat so that the data strategy is aligned with the company strategy and OKRs are in place to execute.
Developing OKRs and then finding alignment between them is not an easy endeavour. It always involves discussion, disagreement, cajoling and testing ideas with your fellow strategists. Although it seems hard, the side effect is that it fosters a spirit of cross-company collaboration and shared mission.
Next steps on the path to execution
Once the OKRs have been outlined and a sense check has passed to confirm that they fully cover the elements of the data strategy, it’s time to move to the more detailed phase of execution.
This involves;
- Developing the metrics and means of collection that will be used to measure progress against each OKR.
- Socialise and fine tune the OKRs with the data teams that will be responsible for delivery.
- The data teams will then determine the tasks and activities needed to make the key results happen. How this happens depends on how your company builds products and services. For example in an agile shop this will then lead to story development and sprint planning.
Conclusion
Having a comprehensive data strategy is essential for modern businesses to thrive. Making the strategy become reality can be tricky. This is where OKRs can help. In this article I’ve shown how a connection can be made between the data strategy and OKRs to build a framework for delivery aligned with the overall company’s direction of travel.
[1] Ron Miller, Big companies are not becoming data-driven fast enough, TechCrunch (2019)
[2] Randy Bean and Thomas H. Davenport, Companies Are Failing in Their Efforts to Become Data-Driven (2019), Harvard Business Review
