Going to play devil's advocate here.
This is one of those "strategies" that looks great in retrospect but doing this up front will be nearly impossible within the confines of a non-fictional organization. This is so highly dependent on multiple critical factors such as resource availability (do you even have enough devs to allocate to this parallel strategy?) and business buy-in (good luck telling the business that you are going to divert all these tech resources in order to throw out all but one of these experiments many months from now). Unfortunately I can't even get companies I work at to agree to 20% time let alone telling them we can't make decisions so we're going to burn huge amounts of cash trying to reach one in parallel.
Like the article said, people are much more cost averse. People includes leadership and product. Both of which will almost always decline this strategy.
>Like the article said, people are much more cost averse. People includes leadership and product. Both of which will almost always decline this strategy.
If you're saying this, then call it like it is: you're calling leadership and product stupid. Because unless you substantively disagree with the logic in the article, you're agreeing that the parallel experimentation strategy does all of:
reach a solution faster,
cost substantially less,
and result in greater wealth generation later.
Later here meaning only "within 5 years," so not even that much later. If you're a business, what else do you care about than wealth generation?
The logic in this article is way too cut and dry in almost every way. For example the article exclaims that "parallel is always faster than serial". That assertion does not hold up in software engineering and it doesn't hold up with teams either.
Some counter points: I'm not sure that parallelizing the experiments would keep the same 4 month implementation time frame. It would also fragment the knowledge base, as fewer people would have been working on the one version that gets pushed. I also have my doubts about what happens when two of the experiments seem equally promising.
Like the article said, people are much more cost averse. People includes leadership and product. Both of which will almost always decline this strategy.