The whole point of putting "Value for the Customer" as the first principle of Lean Thinking is to remind the company leaders that they should constantly fight against that inevitable tendency when organisations scale and become more bureaucratic.
Concrete examples of what this can look like ina. company:
- the "Obeya", a room with a big wall where the Chief Engineer / Chief Product Officer describes what good looks like for their product. It is a way to regularly stimulate those conversations and go deeper on the purpose of the product strategy than "maximise these metrics".
- the "gemba" visits from top leadership, regular visits to actually go and see what teams on the ground are doing, without judgement, to initiate those conversations. "What are you working on? The setting to turn off the vibrations? Interesting, can you tell me more?"
I don't think this can be written by a lot of people even though I think a lot had a similar idea.
Stating that metrics (the detailed, user engagement kind) actually degrades products could be suicidal in tech.
As engineering minded people, it's very comforting to think that we can measure everything and have a number attached to every little thing, when in fact the most important things are very hard or even not possible to measure.
What is your take on metrics - they are fine if we add principles to them? They are good for operational efficiency?
I don't quite agree with the principles thing. I think you're right to say we can't improve a product with metrics but we don't have to set principles which just block any arguments. We just have to have real users test the product and give feedback.
Even if someone comes up with a feature just because they think it will make them look important, that feature - as all others - should be tested (after all other automated tests) in production by users.
Those "tests" will yield results in the form of user feedback. This can also be a metric, albeit a meaningful one. If users are unhappy with a feature, we delete it.
But sadly, I have to confirm the (usually) large corporation culture effect. Changing that is not easy.
I don't understand what you mean by "I don't agree with the principles thing". Testing in production with users isn't some sort of external, infallible truth. It's just as subject to bias & gaming as any other metric.
When I worked on Messenger I remember the waves of "add my little feature to get a 1% lift" followed "oh crap the UI is cluttered I'll get rid of excess features and get a lift of 1%". That's where principles come in--I don't care if that gets you 1% lift, a good review, and a bonus, that's just not the kind of thing we do in this app.
I understand now, yes this seems sensible. I didn't mean to say "anything must be able to go". What I meant was that we can test an idea doesn't work instead of forbidding it by principle. Unfortunately, the extreme of this is that people will push all kinds of crap out, damaging the UX of the product. I can see how that can happen easily on large products like Messenger.
Great post! I'm not totally convinced, though, that the fitbit thing is enshittification. Cory Doctorow defines enshittification as the reallocation of value from users to business customers, and then from business customers to shareholders. That is, it's the intentional exploitation of captive users for financial gain. The mechanism you describe, metrics-gaming, is different, because shareholders do not generally benefit from it — employees do.
I can see that I've stretched the definition a bit. And... the folks in these companies have convinced themselves that individually attributable number-go-up *is* the best way to increase profits.
People’s practices are a consequence of principles. But the principles are a consequence of values. These values can be a sneaky thing to determine. This is leadership comes into play: values.
To discover people’s values, it is interesting to present them with a scenario and ask them to make a choice. Ask them if they’d like more or less of something.
Sometimes you discover they don’t care. That says something about their values too.
For example, once upon a time I built an IDE for a living. We cared a lot about how long it took a user to complete certain use cases or to get feedback. From that came guiding beliefs (principles). These made it really easy for us to make what otherwise could be very hard decisions.
What I was thankful for is that we had leadership that had a strong opinion about what they valued with respect to customer experience. They were vocal about it.
Like a lot of things, there are likely many perspectives here. There could in fact be many reasons why they would add such a feature. I can’t speak to Fitbit, or know if they built that first, but my Apple Watch has that feature and I _appreciate_ it. Now in the case of my watch, it will pop up that notification and requires me to acknowledge. And it’s not just walking, it knows several of my “activities” patterns.
So there could be pressures of maintaining feature parity, but there could also be other reasons such as user requests. Point is we don’t know, so assume good intent. Generally, I would assume they’d appreciate your candid feedback. :)
Very good example.
The whole point of putting "Value for the Customer" as the first principle of Lean Thinking is to remind the company leaders that they should constantly fight against that inevitable tendency when organisations scale and become more bureaucratic.
Concrete examples of what this can look like ina. company:
- the "Obeya", a room with a big wall where the Chief Engineer / Chief Product Officer describes what good looks like for their product. It is a way to regularly stimulate those conversations and go deeper on the purpose of the product strategy than "maximise these metrics".
- the "gemba" visits from top leadership, regular visits to actually go and see what teams on the ground are doing, without judgement, to initiate those conversations. "What are you working on? The setting to turn off the vibrations? Interesting, can you tell me more?"
Thank you for the article.
I don't think this can be written by a lot of people even though I think a lot had a similar idea.
Stating that metrics (the detailed, user engagement kind) actually degrades products could be suicidal in tech.
As engineering minded people, it's very comforting to think that we can measure everything and have a number attached to every little thing, when in fact the most important things are very hard or even not possible to measure.
What is your take on metrics - they are fine if we add principles to them? They are good for operational efficiency?
Good thing I don't know any better.
Metrics are great. Confusing the map with the territory is a mistake.
Excellent article!
I don't quite agree with the principles thing. I think you're right to say we can't improve a product with metrics but we don't have to set principles which just block any arguments. We just have to have real users test the product and give feedback.
Even if someone comes up with a feature just because they think it will make them look important, that feature - as all others - should be tested (after all other automated tests) in production by users.
Those "tests" will yield results in the form of user feedback. This can also be a metric, albeit a meaningful one. If users are unhappy with a feature, we delete it.
But sadly, I have to confirm the (usually) large corporation culture effect. Changing that is not easy.
I don't understand what you mean by "I don't agree with the principles thing". Testing in production with users isn't some sort of external, infallible truth. It's just as subject to bias & gaming as any other metric.
When I worked on Messenger I remember the waves of "add my little feature to get a 1% lift" followed "oh crap the UI is cluttered I'll get rid of excess features and get a lift of 1%". That's where principles come in--I don't care if that gets you 1% lift, a good review, and a bonus, that's just not the kind of thing we do in this app.
I understand now, yes this seems sensible. I didn't mean to say "anything must be able to go". What I meant was that we can test an idea doesn't work instead of forbidding it by principle. Unfortunately, the extreme of this is that people will push all kinds of crap out, damaging the UX of the product. I can see how that can happen easily on large products like Messenger.
Great post! I'm not totally convinced, though, that the fitbit thing is enshittification. Cory Doctorow defines enshittification as the reallocation of value from users to business customers, and then from business customers to shareholders. That is, it's the intentional exploitation of captive users for financial gain. The mechanism you describe, metrics-gaming, is different, because shareholders do not generally benefit from it — employees do.
I can see that I've stretched the definition a bit. And... the folks in these companies have convinced themselves that individually attributable number-go-up *is* the best way to increase profits.
I’m in agreement, but with an addendum.
People’s practices are a consequence of principles. But the principles are a consequence of values. These values can be a sneaky thing to determine. This is leadership comes into play: values.
To discover people’s values, it is interesting to present them with a scenario and ask them to make a choice. Ask them if they’d like more or less of something.
Sometimes you discover they don’t care. That says something about their values too.
For example, once upon a time I built an IDE for a living. We cared a lot about how long it took a user to complete certain use cases or to get feedback. From that came guiding beliefs (principles). These made it really easy for us to make what otherwise could be very hard decisions.
What I was thankful for is that we had leadership that had a strong opinion about what they valued with respect to customer experience. They were vocal about it.
Like a lot of things, there are likely many perspectives here. There could in fact be many reasons why they would add such a feature. I can’t speak to Fitbit, or know if they built that first, but my Apple Watch has that feature and I _appreciate_ it. Now in the case of my watch, it will pop up that notification and requires me to acknowledge. And it’s not just walking, it knows several of my “activities” patterns.
So there could be pressures of maintaining feature parity, but there could also be other reasons such as user requests. Point is we don’t know, so assume good intent. Generally, I would assume they’d appreciate your candid feedback. :)