Five Attributes of Lean Insight & Analytics Teams

If I said I wanted a lean research project, chances are you’d think faster, cheaper and probably digital.

And you’d be right: these are important dynamics in lean insight and analytics teams.

But they’re not the whole story.

When Taiichi Ohno – Godfather of Lean Manufacturing – developed the Toyota Production System, his original focus was improving quality. In the 1950s, Japanese carmakers had a dismal reputation – ill-fitting doors, poor electrics, exhausts falling off on longer journeys. ‘Made in Japan’ was shorthand in Europe and the US for ‘cheap crap’.

Ohno led the design of systems that improved quality dramatically. His legacy of Kaizen (continuous improvement), Kanban (process flow management), and Just-in-Time supply chains are now global standards in high volume manufacturing – and Toyota is the world’s largest auto maker.

In his book The Lean Startup, Eric Ries took Ohno’s principles and added some of his own to apply lean thinking to software businesses. Crucially, he put early and frequent user feedback at the heart of his lean model. Much of this thinking has since spread to mature organisations, and many corporate teams are now urged to think lean as they innovate.

So yes – lean is about speed, value and technology; but it’s also about quality, process innovation and properly embedding the customer’s perspective.

Customer insight and analytics teams have begun to adopt lean principles. Here are five attributes that characterise successful teams.

1 They are confident making decisions using Minimum Viable Evidence

The Lean Startup introduced the concept of The Minimum Viable Product – the idea that is it better to get product into the hands of users as early as possible, rather than refining with a perfectionist ambition and launching too late.

Good enough today is better than perfect tomorrow (by which time competitors will probably have moved too far ahead).

“If you’re not embarrassed by the first launch of your product, you’ve launched too late,” said Reid Hoffman, founder of LinkedIn.

For insight teams, this means getting usable data today rather than building a library of evidence tomorrow.

This can be a challenging shift for many researchers and analysts, but they are facing a demand crunch that goes something like this:

  1. There is now much more customer data available to support decision-making …
  2. … and stakeholders expect more decisions to be underpinned with data …
  3. … so this increases demand on research and analytics teams …
  4. … but headcount in these teams is not growing in line with this demand …
  5. … which means the number of decisions-per-researcher and analyst is increasing.

So – perversely – in a world of more data, the modern insight professional has to be confident making decisions with fewer data points and more expert judgment. It’s a point that Stan Knoops of IFF makes far more eloquently here.

Fortunately, there are now many ways to get your hands on just enough data to help with this high-tempo decision making: sources of free, cheap and big data to analyse search, social and e-commence trends; smartphone qual and ethnography apps for fast contextual insights; and automated survey-based tools to prioritise and validate ideas.

Sometimes, less really is more.

2 They adopt methods and software platforms for iterative insight

The Build-Measure-Learn principle is core to the lean movement: once your MVP is in the hands of users, make sure you capture feedback and performance metrics continuously. Software products generate streams of quantitative usage data; there is also no substitute for qualitative user interviews to work out what’s really going on.

The Build-Measure-Learn approach enables continuous improvement and frequent product updates; many software companies regard their products as living in perpetual beta as they tinker and improve on a weekly release cycle.

Most modern analytics teams get this; most research teams have yet to catch up.

The sequential ‘project’ paradigm still dominates the thinking of research teams and agencies: a business question generates a brief; a project is designed to gather data; that data is analysed and reported (some weeks later). This made sense in the 80s and 90s when market research came of age and data gathering was slow, expensive and vulnerable to small errors.

But it’s not how the world works now.

Modern insight is about frequently testing lots hypotheses: questions that need to be asked directly to people, or experiments and analysis that can be worked with existing data. It’s about regular, small scale test-and-learn – not waiting for a big reveal several weeks down the line. And it’s about building long term data and insight resources – not individual projects.

The tools for iterative insight include optimisation and A/B testing platforms; communities and customer advisory boards; CRM platforms that enable progressive customer profiling; on-going UX and CX feedback programmes; and rapid testing platforms that automate much of the research process.

Lean teams think in terms of iterative insight, not the perfect project.

3 They value qualitative understanding as highly as quantitative measurement

The ‘Five Whys’ is another of the Lean Startup’s central ideas, and another adaptation from lean manufacturing.

We’ve all known – and probably been – that annoying four-year-old who asks whyrepeatedly, and never hears an answer that can’t be met with a further why. It turns out we should encourage that child rather than wish it pain: in complex manufacturing as in software development, an issue buried several ‘whys’ down can cause big headaches and costs that directly affect the finished product.

Here’s an example from Toyota, courtesy of The Lean Startup:

1. Why did the machine stop?

There was an overload and the fuse blew.

2. Why was there an overload?

The bearing was not sufficiently lubricated.

3. Why was it not lubricated sufficiently?

The lubrication pump was not pumping sufficiently.

4. Why was it not pumping sufficiently?

The shaft of the pump was worn and rattling.

5. Why was the shaft worn out?

There was no strainer attached and metal scrap got in.

Laddering questions like this to find deeper answers and motivations – rather accepting the superficial – has long been a staple technique in qualitative research.

In a world where more descriptive data is available to tell us what is happening – which page design converts better, which shade of blue drives more clicks – it becomes even more important to understand why people choose and behave as they do. Insight teams who figure this out give far better upstream support to their innovation and brand colleagues.

The lean toolbox for qualitative insight includes platforms that enable virtual face-to-face dialogue with customers and users – for depth interviews or usability tests; online tools for collaborative brainstorms, forum discussions or focus groups; neuromarketing tools using webcams to decode emotions in facial expressions and track eye movement; and machine learning techniques for making sense of unstructured text, image and video datasets.

4 They combine multiple perspectives and skillsets for integrated insight planning

In lean manufacturing, the concept of Small Batch Flow refers to a counter-intuitive learning about process efficiency: it is often better for one person to do several different tasks in a process than it is to split those tasks between separate specialists.

Ries uses the example of stuffing hundreds of envelopes for a mailing campaign. Conventional wisdom suggests it would be more efficient to separate the letter folding, envelope-stuffing and addressing tasks between three people in ‘specialised’ process roles. In actual fact, the whole job completes faster if each person takes a third of the letters and does all the folding, stuffing and addressing.

The lesson for insight teams is to guard against excessive segregation of skillsets. Large agencies are particularly vulnerable to this, and often generate diseconomies of scale as they add more and more job titles to an account team.

Integrated insight planning is source-agnostic, draws on multiple perspectives and prioritises the commercial and customer questions over the method.

In practical terms, this means building T-shaped skills for insight professionals: balancing depth in one area with breadth across adjacent skills. It also means combining consumer insight teams and analytics teams, or integrating UX research and data science teams as per this example at User Testing, courtesy of Chris Abad, VP Product & Design.

5 They are not afraid to pivot and embrace radical change in tools and methods

In The Lean Startup, willingness to pivot is critical: if you are going to ship product to users early – and commit to learning from them – you need to be open to changing course radically if that early feedback tells you to. This could be as dramatic as junking the entire product roadmap and focusing instead on a minor feature; or it could be a big change to the business model. Eric Ries lists around 20 different types of pivot in the book, but the underlying message is common to all: startups need to expect big shifts in direction.

For established insight teams and agencies, this also holds true. Embracing painful change can be essential; sticking relentlessly to existing methods can be fatal. It’s the Innovator’s Dilemma: what got you here won’t necessarily get you there. Even worse: past innovations and successes can actually create a prison from which it is very tough to escape.

Leading insight and analytics teams are pivoting themselves before their tools and methods become obsolete: replacing laborious CX feedback surveys with lightweight embedded NPS apps; swapping costly brand tracking programmes for social and search lead indicator metrics; dialling back agency partnerships and in-sourcing with software platforms.

Leave a Comment