To quote Chester Karrass, you don't get what you deserve. You get what you negotiate. In this case, what you negotiate with is the "market."

The spectacular products will get found no matter what, and the awful
ones will tank no matter ones, but it is the middle ground of "good but
not spectacular" (GbnS) products where this question is relevant. Every such GbnS startup thinks it is full of ninjas and rockstars that will deliver the next blockbuster, no help needed from those stinkin' marketers.

To paraphrase Prostetnic Vogon Jeltz from The Hitchhiker's Guide, "Apathetic bloody engineers. I have no sympathy at all."

Most GbnS startups DO get EXACTLY what they deserve, which is "killed in the market." Our world has far too many products and services, and losing yet another probably-not-necessary GbnS is no big loss.

And they deserve it because they invest far too little in PR and marketing, far too late, and with far too little discipline and too much reluctance, grudging every dollar spent.

To put numbers on this, they spend between 10-20% what they should on marketing, about 2 years too late on average, and with zero discipline. I have no sympathy for them.

It is the revenge of the marketers/business types against snooty engineers who think "marketing" means slapping a Like button on something and making vague noises about "analytics" (a subject which, despite their greater mathematical sophistication than marketers, engineers spectacularly misunderstand).

Yeah, keep passing those Dilbert jokes around. Would you like fries with them?

For the record, I have 3 engineering degrees, and no business/marketing degrees... just experience, some of it harsh.

And endless respect for those who can do marketing well. I learned late, but I learned (and am continuing to learn).

Which is more than can be said for most engineers. A great marketer is worth at least 10 great developers.

GbnS products get exactly the Darwinian fates they deserve because they simply don't understand what it means to take marketing and PR seriously. Here's the typical trajectory.

1. Month 1: Come up with a "great" (but really GbnS) product idea, spend 6-12 months building it
2. Month 12: "Oh we need a name." A hurried whois search ensues and usually a terrible compromise of a name (not a brand, let alone one with a positioning hypothesis or a narrative) is made up. Of course, they have NO idea that they've done the marketing equivalent of making a strategically critical decision like "LAMP vs. .NET" on the strength of a coin-toss. It goes downhill from here.
3. Month 13: The product is out under its mediocre name, and despite some desperate tweeting, nobody joins. Somebody says, "maybe we need a better About page, and an improved landing page." Something is hastily thrown together. This is an important event, it is the first random act of marketing (RAoM). The pattern of under-investment continues. Nobody realizes that the About and landing pages are about as crucial to the marketing as the database schema design is to the product.
4. Month 14: A slow trickle starts, but things don't really get moving. Somebody says "we need PR" somebody else says "we need user testimonials." More RAoMs follow.
5. Month 15: A few users join and immediately many flaws are exposed. The team goes right back to the trenches for V 2.0. The marketing falls by the wayside.
6. Month 22: V 2.0 is out. Rinse and repeat, same pathology. Somebody comes up with the obvious point, "hmm, our blog and twitter stream kinda went dormant while we were building V 2.0. Maybe we should have kept that going more steadily. Halleluejah! The first glimmer of enlightenment.
7. Month 24: Finally somebody starts asking serious questions about a sustained marketing/PR effort, a carefully thought-through positioning, alignment of messaging, channel selection, and so forth.
8. Month 26: Money runs out.
9. Month 27: Nothing has been learned. Start the cycle again.

Here's the RIGHT way to do GbnS products. (Spectacular products can of course, break any and all rules they like).

1. Start your marketing presence before (ideally 2-3 years before) you even write a LINE of code, when the product is no more than a twinkle in your eye. Just start establishing category thought leadership in a relevant sector and start getting to know the market.
2. When you've got a trusted and well-known voice and a nice thriving and relevant channel of messaging that YOU own (like a blog, or a YouTube channel) with an audience that is clearly characterized, THEN you start thinking of various ideas that might serve it.
3. THEN you put up a minimum-viable product that is no more than an email-list sign-up. If you cant' get at least a 100-200 people to sign up within a week, using your captive channel, you are not ready to launch a GbnS product in the space (and don't kid yourself, I am 99.9% certain that your product is GbnS without knowing ANYTHING more about you, simply because only 0.01% of ideas count as "spectacular.")
4. Keep the drum-beat going. When your first release is out, you should have several hundred people already queued up to try it. Push comes to shove, cut down engineering investment to keep the marketing investment healthy and growing. Even if your product idea tanks, a strong marketing asset can be reused for another idea. A failed technical product? Not so much.

After that... you can fill in the blanks etc. I won't belabor the importance of thinking through positioning, messaging, brand narrative and the like.

So if you've just started a startup, it's already too late to start marketing
If you've just finished a prototype, it is CRITICALLY late to start marketing and I have little hope for you.