I did not find the article very compelling.
The biggest problem is the lack of context. What kind of products is the author talking about? I got the impression that this is mostly about SaaS offerings, but who knows?
Moreover, the arguments presented are insufficient with regards to it's thesis.
The author claims that traffic spikes are actually harmful, but then just argues that they are mostly irrelevant.
The only argument towards actual harmfulness presented is that disappointed users could spread negative sentiment. I don't find that very convincing.
Something conspicuously absent from the article is a discussion of conversion cost. If you have a freemium offering how much do free users actually cost you? Especially for consumer facing software the cost can be really low. It's not like you have a sales team doing calls with each potential customer, people crowding your store.
Might a surge of signups still be profitable, even if the conversion rate is much worse?
But even if you accept that unmitigated spikes are harmful, aren't the remedies easy to apply. If you notice an unwanted spike you can easily increase friction by e.g. limiting new signups to referrals until the spike is over, temporarily restricting your free tier, etc.
In conclusion, traffic spikes might not be that harmful and are easy enough mitigated as to not warrant taking action to prevent them.
My take is that they’re not necessarily harmful in and of themselves, but it’s absolutely harmful to think that this is the way to grow and get traction. It’s not a repeatable approach and it’s likely to pump some top of funnel metrics temporarily without having meaningful impact to the bottom line.
This can be very distracting if you’re pursuing it intentionally and treating ‘going viral’ as a prerequisite to success.
What if you were able to get helpful support, 24/7/365, with no time waiting in a queue, in your own language (regardless of the service provider's location and 'native' language support)? And the company was able to provide the product and support for it cheaper, resulting in less cost to you?
We're far from there, but I expect it'll happen.