Smart, Repeatable, Valuable: From Experimental Mindset to Durable Models Using Lean and Agile

At Synthesis, the consulting problems we work on day to day are as old as brands and marketing (i.e. not very, but somewhat):

• How can we better understand and connect with Growing Audiences?
• How can we shape and adapt to Changing Preferences?

But when we started Synthesis in 2017, we did so because we’d grown dissatisfied with the lack of innovation in the way insights and strategy agencies were working to solve these problems.

We saw a huge opportunity in leveraging Open Data — what people do, what they search for, and what they share online — and in bringing to life the context and culture shaping that data in our work.

One major step was to develop technical solutions with the potential to change the way our partners use open data to understand people.

Travelling to the point of having validated tools which provide our partners with the confidence and clarity to act meant borrowing heavily from the world of Lean & Agile software development.

Consulting x Discovery (Lean)

Both sides of our work — consulting and product development — feed into one another.

The bespoke models we create for our partners in our consulting work lead to us developing reusable tools for working with data, such as our proprietary algorithms and natural language processing models.

These tools form the basis of our client-facing products (launching Q4 2020), which empower our clients to work with open data using Human Centred Data Science themselves.

Every consulting partnership we take on thus represents an opportunity to either enhance the frameworks we have developed or to create something new.

This process requires both us and the partners we work with to embrace an experimental mindset: to work iteratively together until we both feel confident that the models we’ve built provide a robust solution to the initial problem statement.

Often, this means our partners being bold, since they’re buying a set of desired outcomes rather than a fixed, known process. This is where Lean, with its focus on fast iteration, comes into its own.

It’s also why we use the language of partners, not clients. Their efforts to work with us and push what the approaches of Human Centred Data Science are able to deliver have made our work sharper as well as delivered better outcomes for them.

Smart, Repeatable, Valuable (Agile)

Now for the Agile bit. At the end of each consulting deliverable, we run a retrospective meeting in which we ask three questions.

These questions determine which parts of the tech developed in the course of the project will be pushed to our backlog to enhance our internal tools or our external products:

• Is it Smart: did this work deliver something which felt truly distinct and innovative vs. the alternatives which exist for the same marketing problem statement? Did we get to a more honest, more revealing, or more evocative insight on the people behind the data?
• Is it Repeatable: can we effectively systematise the workflows in the process to automate components while ensuring we deliver a consistent, high quality output? What are the variables/biases we had to control for in the data and can we codify those treatments?
• Is it Valuable: did we deliver something which our partner organization wanted to embrace? Did it spark debate, action and change in their decision-making? Did the outcome achieved justify the cost of our work?

Some of these questions we can evaluate and answer internally. But the real proof that we have something worth taking forward to development doesn’t come from those retros.

Measuring success

There are two ultimate measures of success that tell us we’ve hit something exciting (note: neither of them is how much revenue is generated).

The first measure; did we win the second project with that partner? The best confirmation that we’ve delivered something valuable is that our partners want to repeat that work with us in another market / for another brand.

The second; did our partner get promoted? Okay… so this doesn’t happen after every project. Plus, our partners are some of the sharpest people in marketing insights and strategy, so we play only a small part here.

But it’s a good sign that we’re delivering smart solutions when the people who take them forward and implement them are being recognised. We celebrate their promotions as much as any that happen internally.

Conclusion

We’re still in the process of continuously improving and iterating before we bring our products to market. However, the language of Lean and Agile has helped us along our journey of answering traditional consulting problems with technical innovation.

If you’re looking to get from experimental models to executable solutions that are smart, repeatable, and valuable, I hope there’s something useful for you in our experience in making Lean and Agile work for us.