How to set up a Data Innovation Lab

I’ve been working for a few years running a Data Innovation Lab. This kind of entity is still a rara avis, and the vast majority of organizations don’t know how to deal with it. Sometimes the Data Lab emerges because of the “marketing need” created by the Big Data wave: “Of course we are doing Big Data”… Other companies evolve -aka. rename- a part of the existing Business Intelligence department and call it Big Data Lab or Innovation Lab. In certain companies, you can even find both co-existing at the same time.

No matter what the real trigger is, my advice is very clear: set it up properly or don’t even bother… Otherwise you are going to spend more time in justifying your existence than in making a contribution to your company. In this post I want to help you with basics you need to address to set up a Data Innovation Lab

Clearly formulate your Mission Statement

As nobody in your company is going to know why the company needs a Data Innovation Lab, you need a clear answer to the question What are you here for?. For example, our mission statement consisted in three bullet points and a summary statement:

• To create innovative product to maximize the value of the data for your Business.
• To maximize the value of data for Business in [insert company name].
• To make the most of own and foreign data assets to turn  [insert company name] into a data driven company.

We provide innovative solutions based on existing [insert company name] data assets and data assets available in the digital world.

Fulfill the 3 core pre-requisites:

There are three basic things you need when you set up a Data Innovation Lab… If you miss one of them, your life is going to be hell -here speaks the voice of the experience-:

• Mandate (Organizational Acceptance required for us to do our jobs)
• Budget
• Proper team set-up

Getting the Mandate seems to be very easy at the beginning but nailing it down to concrete terms is a hard job. I invite you to go through the checklist below to understand if you have the “mandate”:

• What we do is known and understood in the organization.
• There’s no need to constantly justify that we do what we do.
• If there are other units (e.g.: global units, other internal units) doing similar things, the interfaces and responsibilities are clear.
• We have clear goals our performance is judged upon.
• We get access to all corporate data assets we need to do our job without having to go through any one-off request process.
• No other team is doing the same thing under other name, the same thing under the same name or a different thing under the same name.
• The way of engaging with us is known to the main stakeholders (VPs / Heads level in Marketing, Sales, Customers Services across all B2C, B2B and B2P units)

Budget or funding is the very first question… I’m assuming you got it, otherwise your Lab set up plans can be forever buried. When you request the funding for the Lab, make sure you ask for:

• Allocated budget to guarantee the delivery of our lab projects
• Which is not in competence with other (IT) topics
• Which is adjustable every year (or every 6 months) based on the benefits we provide back to the business

When I speak about a proper team set-up, I mean:

• A team of highly specialized developers, UX engineers and data scientists
• with continuity (intern or long-term contractors)
• with business understanding
• with focus on failing fast rather than building the perfect chocolate teapot

Defend your particular way of working:

We are a LAB and work as such. Our methodology can be defined as Lean innovation with business engagement:

I) Business needs come in with Stakeholders defined and involvement in the project clarified
II) Lean Innovation Cycle

 do {
   1) Ideas are generated around these needs and how we could address them with insights
   2) Data sourcing strategies are defined and implemented in a prototyping way (what's available and we have it, what's available and we could have it, etc)
   3) Basic modeling and insights consolidation is done
   4) Results are exposed back to the business, typically in form of a self-service tool
   5) Evaluation / Quantification of the value generated
   6) Decision: Kill / Iterate one more / Turn into product
 } while (Iterate)

IV) Productization (might consist of a few more iterations to harden the tool)

V) Transition to operations (while keeping the data platform to keep innovating upon)

Understand what you need to succeed (Key Success Factors)

• Multipurpose and multi-stakeholder: data is business units and departments agnostic and the tools we produce cannot be trapped in one silo
• Holistic view: our approach focuses on the cake and not on one of the pieces… otherwise decisions are just partially right
• Autonomy: the product management is driven by business needs but done internally and remains responsibility of the Lab manager to ensure consistency and re-usability
• Measurability first: we are in a position with business to quantify the impact we make

Ger your Targets properly defined

Having targets gives you the credibility you need to drive an innovation function. You can formulate your targets as a Lab in different ways:

• Hard KPIs, for example, based on the quantified impact we have made in the different business areas with our products
• Soft KPIs, for example, based on the acceptance and usage recurrency by the different business units.
• Mixed of both, i.e.: based on number of projects delivered vs. resources required and time employed.

Add-ons to the way you define your Lab you might find useful

1. The Lab is not as a mere analytical capability, as it produces insights also for the Market Intelligence and Market Research functions
2. As a Lab you have the responsibility for the whole stack (from data gathering to insights presentation and reporting)
3. As a Lab you also have the autonomy to engage with business and generate value through innovation
4. You are not  subject to any Request / budget-based business engaging process.
5. You need a channel to present back the results to the top management -otherwise, you might lack the attention you need-

If you are working in the data innovation space, I’m sure you have been faced with situations where you realized that one or many of the aforementioned points have not been properly clarified. I hope this post helps you assessing where you are, providing you with awareness and making the right questions.

Talking about questions, I want to leave you with a difficult one… What is the best place to host the Data Innovation Lab? IT? Marketing? the Business Intelligence Department? Strategy? or just as independent entity? What do you think of turning it into a profit center?