I can't really treat this article seriously when his second graph illustrating "real" growth of startups requires that we are able to breach time-space continuum and go back in time.
Then you wasted an opportunity to learn something from a very insightful article because of a superficial detail. I'm not entirely sure why that's something you're proud enough of to post a comment about it. Seems like a flawed outlook on life to discard a message because of a single superficial detail in its delivery.
If everyone had infinite time to read articles on the internet, it would be best to keep reading and possibly learn something useful from the article despite its error. Since we don’t have that much time, we have to decide which articles are worth reading – there is a lot of nonsense on the internet that is a waste of time to read (Sturgeon's law). Errors of sufficient number and magnitude at the beginning of the article indicate a higher likelihood that the rest of the article is too nonsensical to bother reading, i.e. that the ratio of usefulness to time spent is too low.
Now, maybe you disagree that a graph with loops is that big of an error. But then you should write about the graph specifically. I don’t think you have grounds for disagreement with the basic principle of generalizing the quality of an article from part of it and using that estimate to decide whether to continue spending your time on reading that article.
It doesn't represent data or a function - its just an illustration that borrowed from graph lexicon. Had it been styled as 'serious' graph it would have actually been misleading.
I wouldn't bother about loops, they are allowed if you're not graphing functions[1] and are quite common on (x1/x2) graphs as opposed to (x/t) graphs.
My real issue with it are unlabeled axis - without them those graphs have no meaning except as an conceptual illustration of an idea that something loops around while growing. Also, OP assumed that x is time, which it does not have to be for the "graphs" to make sense.
Yeah, yeah, his graphs suck. His point is that good companies tend to go through one or more iterations that are just wrong on the way to something good. That is basically true.
I thought the "graphs" were a good illustration of ups and downs combined with design iterations and user gain/loss cycles. Of course they were abstract but they got the point across.
Sorry for the confusion. I left off the axis labels because the graph was merely a conceptual contrast to the tired hockey stick curve.The idea of its iterative shape, relative to a hockey stick, is the important point.
But I actually like your point because its interesting (a little) to think about what the axes have to be for those shapes to work.
There is probably no confusion around the y axis representing traction - whatever metric is meaningful for the business at that stage. It could be users, or revenue, or engagement, etc.
That pesky x axis is the issue. In the second graph, it cannot be time, as many have pointed out. And that's confusing because the x axis in the first graph of the hockey stick shape is usually time. The x axis in the second graph could instead be a measure of learning, or perhaps product development as one commenter suggested. It probably works best as degree of product/market fit. As prod/mkt fit increases with iterations, so does traction.
I read a little way into the article to see whether he'd explain the lack of axis labels or just how the graph maps to anything real. Nope.
The article is reasonable but the loops just annoyed me too much.
I had the same thought at the first glance at the loopy graph, but decided to take off my nerd hat and put on my "I'm listening to a business guy" hat. I think this is something we tech people tend to get hung up on too much... snickering at the slightly misused technical jargon and other similar errors (like this graph) that business people make and then not being able to hear what the person is actually saying.
Any non-technical, non-mathmatical person would look at that graph and say "yeah, you go around in circles for a while and then finally get things worked out" and I actually thought this was a pretty good post otherwise.
"He was saying a real startup trajectory might look something more like this" (emphasis mine)
When we're graphing trajectories, not functions, loops are certainly allowed (see e.g. phase diagrams[1]). Still, without labeled axis those graphs are pretty much meaningless, so I treat them more like an conceptual illustration.
Sadly I was the same, they lost me at that second graph. Even if I was to excuse the ability to go back in time, I just don't believe that really illustrates how failed products go (ie. very strong growth and then just as fast decline), most wouldn't have the level of traction indicated there.
