Realpolitik of Using Lean Startup in the Enterprise

Posted by on December 19, 2013 in Blog | Comments Off on Realpolitik of Using Lean Startup in the Enterprise

Realpolitik of Using Lean Startup in the Enterprise

Lean Startup – Narrative, Movement, Fad

Is Lean Just Another Fad?As with all meta-narratives, Lean Startup¬†creates all kinds of responses from the business community because many of us feel part of that narrative. “Lean” as a concept doesn’t belong to anyone, including Eric Ries, because its just part of a wider set of narratives that many of us have been acting out for some time, to varying degrees. Indeed, Eric’s site refers to it as a movement. There’s a growing opinion that it’s just a fad.

As an example of the kinds of response, Andreas Klinger recently posted an article with the title “Why Lean Startup Sucks For Startups.

In the most part, he’s not really discussing the core of the Lean Startup narrative at all, but mostly our response to it – namely that once it becomes a process and no longer a tool that is couched within critical thinking, it kind of becomes useless. This is a general pattern in life, so I can’t really argue with it. However, he went on to mention that it might not be all that bad for enterprises. As I’ve tried to extend the narrative inside of enterprises, I felt I had something useful to say.

I try to avoid arguments that are merely my opinion about a concept because what’s the point? We’d have to first agree on what the concept really is, but given it’s mostly a narrative about product management, we would get stuck on semantics. Also, why should I think my opinion is all that useful?

On the other hand, what I can do is to say what I’ve done and try to share some insights, perhaps draw a few conclusions.

What follows is an unedited copy of¬†my post on LinkedIN, in response to a thread discussing Andreas’s blog post.

Lean Startup in The Enterprise – The Realpolitik

I’ve started 5 “Labs” inside of enterprises – i.e. business units with a remit to try new product ideas that are somehow “outside” of the product roadmap and existing processes. The biggest return I’ve had from using “Lean startup” “methods” inside of an enterprise is mostly as a rubric for trying something new that ostensibly has “Silicon Valley” smarts baked into it. In other words, “new and shiny” has its uses.

However, new for the sake of new isn’t that useful. So, the selling point is usually – “we can do this [get traction for a new idea] faster and cheaper [than the current process].” This resonates, especially with board members (and you will typically need one of them to sponsor a new approach), because of too many failed product attempts that cost gazzillions.

The number one threat to a new product idea in an enterprise is death-by-co-option. This is well researched and documented by Clayten Christensen. Enterprise “intrapreneurs” all know this and will do anything to avoid co-opting 1001 opinions from stakeholders and the janitor (everyone will offer an opinion).

How does Lean help?

Well, firstly – and most powerfully (but only works once) – because it is a “new concept,” existing stakeholders don’t really understand it. So long as there’s a senior-enough sponsor for using Lean (whatever it is), this lack-of-stakeholder-knowledge vacuum can be exploited to give valuable headroom to get stuff done before folks insist on co-opting.

Second, and more in line with the method itself, early testing can be used to provide some kind of “evidence” that a new and usually “radical” (by current product standards) idea has some merit. This can work well because the loudest co-opting voices will usually mention why something won’t work – “your assumptions are wrong” – and, to be honest, for good reason based on the standard “voice of the customer” insights available to the business. If early testing confirms a key assumption, then all well and good. You buy yourself more time to innovate.

In reality, there are all kinds of problems with early testing, related to cognitives biases – e.g. overconfidence, attention blindness, hypotheses fallacies etc. As the author points out, it is easy to avoid self-delusion when something clearly doesn’t work, but much harder when something does “work” – or, rather, doesn’t clearly fail. It is very hard (impossible?) to scale qualitative tests to quantitative future performance. For these reasons, I would not trust myself to rely only upon this approach for a real start-up. Critical thinking (the “tool kit” approach) should prevail if it’s your future and money on the line. In an enterprise, critical thinking is a rare luxury and so you have to resort to other (political) means to advance an innovation cause.

In all attempts to use Lean Startup in an enterprise, I have found that it only “works” when trying out new ideas that have yet to be adopted by the stakeholders as actual products. In other words, the results of experimentation mean nothing to them. Once the results do, I have found that it is very hard to use Lean on a real product development. This is usually due to the “horse and cart” effect mentioned by Ford (“they would have asked for faster horse and cart”). What happens is that the concept of “minimally viable” is out of sync with existing product marketing mindsets where brand, current feature sets and all that “baggage” is considered minimally viable because that’s what people have responded to in getting our business to where it is today.

The worst aspect of using Lean Startup in an enterprise is the psychology of “fail fast”. No one likes failure and we prefer to defer pain. Stakeholders get impatient that things are taking too long because of experiments. Early failures are seen as a bad signal, not a good one. Somehow, they prefer to press on quickly, even if it means increasing the chance of failure at the end. Overconfidence illusions are hard to overcome.

Subscribe to Paul Golding's posts and other provocative content by email