3 ms·

> smallest bit that would possibly work Nice thought - but I've tried that in the past, though, and smashed into unreasonable expectations from the clients the

> smallest bit that would possibly work Nice thought - but I've tried that in the past, though, and smashed into unreasonable expectations from the clients themselves: "Why just do the smallest bit? We need the whole thing!" This only works if everybody understands the value, and this shared understanding is rare, to say the least.

jschwartzi 7y ago
Castle walls are built a brick at a time.
- AnimalMuppet 7y ago
  True. But one brick is not a viable castle wall.
  - kaens 7y ago
    good thing no one is advocating for laying one brick and stopping
    - mrmonkeyman 7y ago
      You do need to demo the one-brick wall and convince the suits it will work out in the end. You do this for every brick.
  - breeny592 7y ago
    But it's the difference between building the smallest wall to completion first versus going the whole way around the outside layer by layer. Showing one thing that is complete is better than partial progress everywhere.
mixmastamyk 7y ago
This is why you confirm progress with the client at every sprint/milestone. Unless they expect a full system built by the end of the first, this shouldn't be a problem. If they don't understand project management, our job is to educate them. If they can't be educated for some reason, and there other clients, let them go.
davidivadavid 7y ago
Another source of such mistakes is having too small a team tackling too big a problem, and being unable to ship a solution on time, because you tackled the wrong problem. That's typical when you approach a fairly mature category where the feature set is pretty entrenched, and people have very set expectations. To them, every feature is a must have. The lesson here is that.. you probably shouldn't ever do that unless you have a big enough differentiation that allows you to ship just that differentiation and delivers outsized ROI for the customer, without building all the other expected stuff. In other words, you need to unbundle the part of the value chain where you can deliver the most leverage and refactor everything around that — and not commit the biggest mistake which is rebuilding the whole thing expecting that you're going to add your "killer feature" as the cherry on top. You'll never get there.
hinkley 7y ago
Or the converse, where they want you to stop when the product is kind of half-assed. The pays the same, but the sense of accomplishment goes out the window.

jschwartzi 7y ago

Castle walls are built a brick at a time.

AnimalMuppet 7y ago
True. But one brick is not a viable castle wall.
- kaens 7y ago
  good thing no one is advocating for laying one brick and stopping
  - mrmonkeyman 7y ago
    You do need to demo the one-brick wall and convince the suits it will work out in the end. You do this for every brick.
- breeny592 7y ago
  But it's the difference between building the smallest wall to completion first versus going the whole way around the outside layer by layer. Showing one thing that is complete is better than partial progress everywhere.

mixmastamyk 7y ago

This is why you confirm progress with the client at every sprint/milestone. Unless they expect a full system built by the end of the first, this shouldn't be a problem. If they don't understand project management, our job is to educate them. If they can't be educated for some reason, and there other clients, let them go.

davidivadavid 7y ago

Another source of such mistakes is having too small a team tackling too big a problem, and being unable to ship a solution on time, because you tackled the wrong problem. That's typical when you approach a fairly mature category where the feature set is pretty entrenched, and people have very set expectations. To them, every feature is a must have. The lesson here is that.. you probably shouldn't ever do that unless you have a big enough differentiation that allows you to ship just that differentiation and delivers outsized ROI for the customer, without building all the other expected stuff. In other words, you need to unbundle the part of the value chain where you can deliver the most leverage and refactor everything around that — and not commit the biggest mistake which is rebuilding the whole thing expecting that you're going to add your "killer feature" as the cherry on top. You'll never get there.

hinkley 7y ago

Or the converse, where they want you to stop when the product is kind of half-assed. The pays the same, but the sense of accomplishment goes out the window.