How to embrace Lean Start-up at scale

“Don’t worry about failure; you only have to be right once.” — Drew Houston, co-founder and CEO of Dropbox


  • Start-up behaviours are poorly understood in large organisations
  • Lean is misunderstood, leading to organisational resistance
  • To succeed you must know the levers to aim for ; MVP, A/B testing; test learn and fail fast, standardise, customers
  • Avoid traps; vanity metrics, over-architecting, existing systems; customers; use innovation accounting

There is an almost automatic barrier that goes up in almost any existing organisation when it’s suggested that start-ups have something to offer large organisations. “We’re special! We are too mature an organisation. System XYZ is specially…”

Each one of these rationalised answers is one that should, with a better understanding, be a reason to look to lean start-up to begin learning from, and this stems from a lack of understanding of what defines a start-up and the misinterpretation of how to define lean.

“We cannot become what we want to be by remaining what we are.” — Max DePree

A start-up by its very nature probably doesn’t know what is going to define their long term value. They must find this out for themselves— by failing fast a start-up can move on and learn as it goes, if they fail slowly by relying on vanity metrics (such as hits on a website) then they inevitably disappear. The true benefit of a start-up is that it is a scientific experiment; If it succeeds in growing sustainably then the hypothesis in the start-ups business plan has been validated; if not then the hypothesis has been proven wrong.

The poor and corrupted view of Lean is derived from organisations using it as an excuse to cut costs (or as I have heard it described ‘capacity creation’) without properly embedding Lean principles (customer value, Flow, pull, perfection, and the lessons taken from Toyota by addressing waste (“Muda”), continuous improvement (“Kaizan”) and the ability to stop were where there is a systematic problem (“Jidoka”) and active observation (“Gemba”).
[Readers note : I only use the Japanese words as the English translations are very approximate. You should take time to understand these when you are looking at adopting lean approaches in your organisation.]

Another challenge lean faces in adoption is the blended view that many take in mixing six sigma and lean. While complimentary disciplines, six sigma is borne out of lean thinking and leverages statistical and methodical approaches, with the ultimate aim of mistake proofing (“poke-yoke”, inadvertent error prevention) a process and reducing the time a process takes to maximise the customer value from a process. While extremely effective, six sigma is best applied to existing processes as a continuous improvement toolset.

“If you define the problem correctly, you almost have the solution.” — Steve Jobs

How to succeed. To succeed , you must first know what success is, and you should be willing to change what this means for you. This may seem contrary, and in a slow moving legacy organisation this will inevitably be proven to be the case.

The only way to learn what customers value is to get the product in their hands and try to separate your biases from what they tell you. It is much easier to say the customer is wrong, that admit that your product has a defect. If that product has taken 6 months or a year to get into the hands of a customer, your ability to adsorb that mistake becomes emotionally and financially much harder to take. By taking agile approaches to delivery, be that software, a service or a product you can reduce the financial and emotional impact if you are on the wrong trajectory and change. If it has taken you a year to get there you may not be able to change the product as it will be too late and the financial controllers, business head or business leaders will ultimately pull the plug. This causes so much wasted and ultimately damage that there can be negative feedback that propagates outside of the delivery team itself.

Use 5 Whys and Cause and Effect Diagrams to try and understand what is preventing you releasing an MVP and prioritise the things that are blocking you by plotting these on an Eisenhower Matrix (importance and urgency are on the 2 scales. )

Now that you have a path to release an MVP and are able to prove your product has some customer value what about A/B testing? Why? With any form of A/B testing you can set a hypothesis and test it with groups of customers. Organisations like Amazon, Skyscanner, Google and Microsoft are able to do this at such scale, their practices are known as Scientific Engineering. They are testing many hypotheses at once. At Amazons current release schedule, updates are made every few seconds. By testing testing this often Amazon is able to zone in on customer behaviours and focus on true customer value.

By robustly and frequently undertaking A/B tests you can fail fast. Failure is engrained at a young age as being bad and this is amplified through many working lives where it’s seen on RAID logs, audit reports and company reports. What we forget, is that only by failure do we learn, individually and collectively; the closer in proximity to us that failure occurs, the more we learn from it. The faster we fail, the more capable we are to move on from it.

By standardising approaches that succeed we can reinforce the path to success. Standardisation takes many forms, be that training, processes, systems used. You should allow for standards to be challenged and enable innovations to disrupt those standards set out;

Through all of these steps there should be one thing at the core, your customer. A customer can take many forms and you should not constrain yourself to a view of just one type. These can be stakeholders, users, sponsors and operators. While only 1 of these may generate revenue all can influence the success of the product. You should be mindful of setting expectations for those customers you can influence. Where perfection is expected, nobody will be happy.

This empathy early adopters have is driven out of a 2 way exchange. By understanding the psyche of early adopters you can make best use of this exchange.

We are dying from overthinking. We are slowly killing ourselves by thinking about everything. Think. Think. Think. You can never trust the human mind anyway. It’s a death trap.” — Anthony Hopkins

In the pursuit of adopting lean start-up approaches there are many traps and false wins that can fool you into declaring success. In the early era of the .com boom there was a railroad mentality of “Build it and they will come”. This hype led to many unusual and poor business decisions and as humans we are not immune to this but we can interrupt it. False flags of success such as number of hits on a website or the time spent using a product. Both are typically not qualified and for different reasons. You should check that measures how have are critical to success, actionable and sustainable. The only way to understand this is by checking back with your customer value and business value. You may also attribute growth in use to something you have not factored (a false hypothesis) and only true independent A/B testing will be able to root this out.

A trap large organisations often fall into is over architecting their solution. I myself saw this with a previous employer, that was reluctant to release anything that was not perfect on a range of measures. While for many organisations now, data security is not optional, many other architectural elements can be with early adopters. They may not expect resilience or reliability if they know they are in the early adopter group or even a slick UI.

Often an organisation will look to what they know already; Be that using existing procurement channels, systems they already have or people and teams already in place. This is an institutionalised form of resistance that both gives comfort to decision makers and the system itself. This results in constrained thinking, poor quality uncompetitive solutions and sustains the existing system that further holds back the company or organisation. Accountants have a good way of dealing with this by not considering sunken costs in making financial decisions. You should adopt Innovation Accounting’ within your team to really understand what innovation is driving and what existing systems are constraining. By turning an existing system into an innovation cost you can move forward in your lean start-up adoption

Becoming too reliant on early adopters can, itself, be a trap. For the same reasons that early adopter groups are valuable, they also can lead to a false sense of security in the value your product gives and the quality that can be accepted. You should learn to recognise as your product matures and changes not to be too reliant on this group and simultaneously recognise the value they offer.

“The only way to win is to learn faster than anyone else.” — Eric Reis, author, The lean start-up

In this article I have drawn out key learnings from Lean, Agile and start-up that can be applied. Not all of these are easy to adopt and embed. Organisations are systems, they have resilience that resists change. This is how they have survived and is the same reason they may fail; failing to adopt to changing customer needs and different customer groups; by treating every failure as a reprimand and not using these failures as true lessons to learn from; making expensive long lasting investments with no way to validate early on that these are built on the right assumptions. Tools such as innovation accounting and 5 Whys re useful in progressing forwards.

By building on these and principles from lean, agile and the ones you set out for your organisation, the work you do in an organisation can successful turn nudge and, in turn, direct the systems towards greater agility.