What We've Learnt by Running a Data Strategy Consulting Company
As a Data Science and AI strategic consultancy, we've seen many companies implementing technical solutions that eventually failed to be executed or adopted in the organizations. These companies come to us with different questions: from why it happened to how to make them work. Our 10+ years of experience in strategic advising in the area of data tells us that the only reason for this is the lack of an initially developed Data Strategy that would help businesses oversee different opportunities and make the right choice based on the company's needs.
What is Data Strategy
Data Strategy is a roadmap that shows you how, when, and where to apply data to your business. Having a Data Strategy greatly increases your chances of success with your data initiatives and ensures that your investments will pay off.
The objective of Data Strategy is to deliver a massive positive impact through data, now and in the years to come, considering the changes that might happen in the market over time.
To ensure that your Data Science and Artificial Intelligence projects will be successful, you need alignment across your company.
Organizations come to us too late, not too soon
“We’re not ready for this yet”, “We want to get our data in order first” and “We’re just trying it out internally”.
We’ve heard this too many times. Organizations that are not data-driven, and never were, think they can be successful with data science and AI projects by themselves. Sure, it happens, but most times, data science projects fail to deliver the expectations.
To these companies, we say: Take your data science seriously. Be professional about it. Using your data effectively takes specialized technical and business people. It’s extremely hard to be successful in the first years by randomly choosing an analysis that seems more or less relevant and just pushing a data scientist and a developer to do it. It won’t work.
What happens is many of these companies that were initially reluctant to define a data strategy, end up doing one of two things:
they end up realizing how critical data strategy is and eventually will look for someone who has done it before.
they end up struggling for months on end with analysis that failed to deliver any kind of results only to take years to do what could have been done in months, or simply hiring someone who has done it before, months or years too late.
There is no strategic support
Most companies, when they come to us, have this idea of becoming “data-driven”. This is usually well-intentioned. However, many think that being data-driven is purely a technical challenge, instead of a strategic one.
Becoming data-driven needs to come from the top levels of the company and trickle down to the bottom. This means that before the company starts measuring detailed KPIs at the operational level, it needs to start measuring strategic-level objectives and performance levels using OKRs, or strategic KPIs.
Many CEOs want to know the efficiency of their employees, but will close their eyes and procrastinate when it comes to making their objectives clear and data-driven to the organization. This is a recipe for failure and constant battles between a level that needs to be data-driven and measurable and a higher level that refuses to apply it to itself.
Alignment is the Key to Developing a Successful Data Science Project
By lacking measurable strategic objectives it becomes very difficult to create alignment between the data initiatives and where the company wants to go.
Alignment happens when a company works like a single organism and everyone is heading toward the same goal. The success of all your data projects comes down to shared, well-communicated values within the company.
Typically, when new technology initiatives are introduced to an organization, people tend to misunderstand them. They might be afraid and because of this, reject the changes. What happens after - solutions are developed but not used in the company. The lack of organizational alignment is the number one reason why most Data Science and AI projects fail to be executed or adopted.
Other times, the proposed initiatives don’t actually focus on any bottlenecks. Guess what happens? Once the solution is deployed, there is no change in the performance of the company.
To avoid this, Data Science and Artificial Intelligence solutions should be centered around people's needs and the company's goals and integrate perfectly with the existing solutions. People need to understand why these changes are happening and how technologies will help them personally and the organization itself.
For companies looking to grow, the ultimate objective of a Data Strategy is not to replace people in the companies but to help them do their jobs more effectively and efficiently than ever before. People will be replaced in their repetitive, simple processes. However, once they are automated, people can focus on high-quality problems, problems that only humans can solve, and problems that are truly complex and require creative solutions or human interaction.
Data Strategy makes the way to becoming a data-driven company transparent and consistent. It secures you against possible risks and provides you with a clear picture of how to scale your business with technology.
Most common problems
When a company finally decides to take data strategy seriously, these are the most common problems found:
Foundation:
Lack of OKRs or strategic KPIs: the company doesn’t know where it is, where it wants to go, and if it’s going in the right direction.
Lack of a single source of truth: the data is scattered across the organization, many times living in silos.
Lack of technical infrastructure: there’s no infrastructure to support the company in becoming data-driven.
Observation:
No dashboards: the company doesn’t have any dashboards or visualizations to help understand patterns in the data;
No alerts: the company lacks systems that alert when a target is hit when the performance drops a certain level, or an anomaly is identified.
Resilience:
Subsystems dependency: systems in the company might be over-dependent on other systems, in a way that if one of those systems fails, the whole company might fail. One example is invoicing, many companies will stop if their single invoicing system fails.
Competence:
Processes aren’t clearly defined: many times organizations want to use AI and data science to improve parts of processes that aren’t even formally defined or documented yet.
Expansion:
Jumping ahead: imagine on your first day of running wanting to try a marathon. Many companies are trying to jump straight into the most advanced analysis without taking care of the simple things first. They want to develop and deploy full-scale recommendation systems or super-advanced language models, however, they lack the basic foundation that will support the development, deployment and maintenance of those systems.
Did you notice that, before each of these systems, there’s a category? Foundation, Observation, Resilience, Competence, and Expansion - or FORCE - is our data strategy methodology that tackles each of these challenges and sets up any organization on a successful path to success.
If you’re interested in learning more about how to make your company data-driven, you can read more about it on our dedicated page, or talk to us.