Very interesting. The idea: build something small, measure what clients actually do with it, learn, adjust. Key points:
Learn what clients want
- Observe what actually has an impact on clients, don’t assume. A/B test the website (show two versions, keep the one that works), analyze behavior, measure the % of visitors that take the service. Don’t over-plan or over-engineer on wrong assumptions of what ppl want from your app.
- The common issue: teams are efficient, but at making smth the clients don’t want! Finding what they want is the real goal, not making things.
- Make this build -> measure -> learn loop faster and smoother.
- Try at small scale, even as a huge company, and test it in the real world.
- Be ready to change direction when it doesn’t go as planned.
Work in small batches
- Don’t have ppl stuck with too many tasks at once (the long-term work gets left behind) or seeing only their one part of the product (looks like productivity, but if nobody uses it, what for?).
- Don’t just cut development times: give the engineers the tools to ship in smaller pieces, and explain the company’s situation to them.
Fix the system, not the people
- Often the root of a problem is human -> train the people.
- The 5 whys: when something goes wrong, ask “why?” five times in a row to get from the symptom to the root cause. Put in effort proportionate to the problem.
- Don’t blame ppl, blame the system. Ppl make mistakes, the system should be resilient.
- Start with a small problem, one at a time, and name someone responsible for the 5 whys to lead the meetings.
Innovate inside a big company
- Split into small teams and make them care (shares, fame, authority), not only through money.
- Protect the innovation teams from the rest of the company by giving them independence, but watch their indicators and limit them in time and to a small customer base.
- Let them grow and experiment more. Once the new thing is profitable, the company will try to protect the status quo again.
- Encourage new ideas, but ask what evidence they’re based on.
Grow once you’re successful
The part on growth is very relevant. Where growth comes from, per the book (page below, French edition): word of mouth, ppl seeing the product used, ads paid for out of revenue, and repeat purchase or use.
Ways to grow: keep the clients you have (Pepsi) vs clients bringing in new ones (Hotmail) vs idk
The growth engine is the most important thing. And growth doesn’t prove your new features work: measure them on their own.
private