TRANSCRIPCIÓN 🔗
WEBVTT
00:00:00.000 --> 00:00:02.479
Here's a graph that I have shown many times before.
00:00:02.640 --> 00:00:04.959
In trainings, during conference talks and elsewhere.
00:00:05.120 --> 00:00:09.039
It shows how the marginal cost of software features changes over time.
00:00:09.279 --> 00:00:11.359
Uh kind of, as we'll see later.
00:00:11.679 --> 00:00:15.679
And a talk by Kent Beck inspired me to think about it again.
00:00:15.839 --> 00:00:21.199
The graph below shows the marginal cost of software features and how it changes over time.
00:00:21.440 --> 00:00:26.320
When we start a new project, on the mythical green field, adding new features is cheap.
00:00:26.480 --> 00:00:29.679
But as the months go by, the more expensive it gets.
00:00:29.760 --> 00:00:32.000
And at some point, we seem to hit the wall.
00:00:32.159 --> 00:00:38.320
The cost for adding new features rises so quickly that we barely get any software out to our users at all.
00:00:38.560 --> 00:00:41.840
But the time axis here cannot really be about time.
00:00:42.000 --> 00:00:48.880
When we just leave the software alone, or when it is really done, adding new features does not get more expensive anymore.
00:00:49.119 --> 00:00:56.719
It looks like adding features is getting more expensive over time, but actually it's getting more expensive the more changes we make.
00:00:56.960 --> 00:01:12.640
What I have shown here is the trajectory of a typical waterfall or water scrumfall, also known as Agile Theater Project, where we are always under pressure to deliver and we spend most of our time adding new features and somehow paint over the problems that arise.
00:01:12.799 --> 00:01:16.159
But we could also make some changes to bring the costs down again.
00:01:16.319 --> 00:01:19.439
I can draw a second curve that shows this scenario.
00:01:19.680 --> 00:01:25.280
Every time the costs rise, we make some changes to bring them down again and not to add new features.
00:01:25.439 --> 00:01:32.239
At first, this is more expensive because we are not delivering valuable software features while we work on bringing the costs down.
00:01:32.400 --> 00:01:35.439
But in the long run, it prevents us from hitting the wall.
00:01:35.680 --> 00:01:39.040
Or at least the wall is much further away here.
00:01:39.280 --> 00:01:46.879
Now here's another chart, my recreation of a graph that Kent Beck drew during his conference talk, Software Engineering After AI.
00:01:47.040 --> 00:01:51.599
It is related to what I described before, but not exactly the same.
00:01:51.920 --> 00:02:00.640
On the y-axis, we have futures, also known as options, the different things that we can still do in the future, or the options that we have left.
00:02:00.959 --> 00:02:04.400
The x-axis shows the features our product has.
00:02:04.719 --> 00:02:09.120
We start with zero features, but at that point we have a lot of options.
00:02:09.360 --> 00:02:13.439
As we add new features, we end up with fewer and fewer futures.
Load full transcript
00:02:13.520 --> 00:02:19.360
And at some point we have used up all our options and cannot add any new features anymore.
00:02:19.599 --> 00:02:22.479
We have hit the metaphorical wall from before.
00:02:22.719 --> 00:02:27.280
But just as before, we can work on adding new features every now and then.
00:02:27.439 --> 00:02:32.960
If we get this right, we might postpone a point where we hit the wall, maybe even indefinitely.
00:02:33.120 --> 00:02:39.599
Or even better, we might manage to achieve a trajectory where the number of our futures rises over time.
00:02:39.840 --> 00:02:43.439
Speaking of time, note that this graph doesn't contain it.
00:02:43.599 --> 00:02:48.960
Time is implicit as we follow the lines, every new point on that line is a later point in time.
00:02:49.039 --> 00:02:56.639
But we can move freely along the axis, we can also remove features and increase or decrease our futures by doing so.
00:02:56.879 --> 00:03:01.759
The first graph, the one I have been using in the past, gives us a hint about time and effort.
00:03:01.919 --> 00:03:05.520
We need to make changes to bring the cost down, and thus time passes.
00:03:05.680 --> 00:03:10.800
Also, cost is maybe more tangible for many people working in software development than futures.
00:03:10.960 --> 00:03:14.800
But I also had a Eureka moment when I watched Kent Beck's talk on YouTube.
00:03:14.960 --> 00:03:20.080
The options we have left are a part of the value of our product, and so are the features.
00:03:20.240 --> 00:03:27.759
Still having the possibility to decide has a value on its own, and increasing our options brings down the cost of future development.
00:03:28.000 --> 00:03:33.039
So in Kent Beck's visualization, both axes tell us something about the value of the software.
00:03:33.199 --> 00:03:38.400
By the way, if you want to read in depth about options and value, try to get a copy of the book commitment.
00:03:38.560 --> 00:03:40.639
All the links are in the description below.
00:03:40.879 --> 00:03:41.840
The two are linked.
00:03:42.000 --> 00:03:46.800
Improving our features or adding options brings down the cost of new features.
00:03:46.960 --> 00:03:52.479
Or in other words, the cost of features is a proxy measure for how many options we have left.
00:03:52.719 --> 00:04:00.400
What both graphs don't tell you is how much you lose in the short term by working on adding options and bringing the marginal cost of features down.
00:04:00.639 --> 00:04:03.919
While we are working on that, we won't deliver any new features.
00:04:04.080 --> 00:04:13.120
And if we assume that features created is a proxy for value created, then you create less visible value in the short term, but more in the long term.
00:04:13.280 --> 00:04:15.919
And that is not exactly true for every feature.
00:04:16.079 --> 00:04:20.720
Some do not bring any value, some others might even decrease the value of your product.
00:04:20.879 --> 00:04:23.839
But on average, your features should be valuable.
00:04:24.000 --> 00:04:27.360
I have been thinking about different ways to show that for some time.
00:04:27.600 --> 00:04:32.240
If I ever come up with a good visualization, I will write a follow-up post here.
00:04:32.399 --> 00:04:39.600
And also, the graphs don't tell you why it's definitely a good idea to split those phases to increase options or to add features.
00:04:39.759 --> 00:04:44.079
If you want to know more about that, can't Backs Book TidyFirst will get you started.
00:04:44.319 --> 00:04:49.519
I have worked with clients where it was normal that some things would take forever and nobody questioned it.
00:04:49.680 --> 00:04:53.600
The problem is, once you're at this point, nothing is cheap or easy anymore.
00:04:53.759 --> 00:05:00.639
Going forward, doing business as usual is expensive, but trying to climb down from that mountain is expensive and hard.
00:05:00.800 --> 00:05:03.360
If you're not there yet, look at those graphs.
00:05:03.519 --> 00:05:05.519
Try to assess honestly where you are.
00:05:05.680 --> 00:05:10.000
Think about what you can do to bring the marginal cost down just a little bit.
00:05:10.079 --> 00:05:12.480
And then prioritize it and start working.
00:05:12.639 --> 00:05:24.000
Oh, and if your response is we use AI, this does not matter to us anymore, I am very sure that working on increasing options and keeping feature costs down still matters, or maybe even matters more than ever.
00:05:24.160 --> 00:05:26.480
But this is also a topic for a future, I say.
00:05:26.879 --> 00:05:34.160
Oh, and also in the first graph, I explained that the x-axis cannot be about time, but it also cannot simply be about changes.
00:05:34.319 --> 00:05:38.240
As the software becomes more complicated, the frequency of changes goes down.
00:05:38.399 --> 00:05:43.759
So it's time again, but only as long as we make changes, or something like change time.
00:05:44.079 --> 00:05:47.519
If I ever come up with a good explanation, I will write it up here.
00:05:47.680 --> 00:05:52.879
But until then, or until my next blog post, have fun designing and developing software.
00:05:53.040 --> 00:05:57.680
And if you have learned something, or if you have any questions, tell me in the comments.