Send a simple, one-line email that asks them what they hope to achieve with your product/service.
"Hi, saw you signed up for Acme, welcome! What are you hoping to get out of your trial?"
(I've written about this before: http://www.gkogan.co/blog/question-for-saas-trial-users/)
What has led me to submit feedback most often is having a fairly prominent "send feedback" button somewhere on every page. It means as soon as I notice a problem or think of a useful feature, I'm immediately able to press a button and write even just a one sentence suggestion.
Some confirmations that your company will deliver the whatever Acme supposedly provides according to what your PR and (insert here either "really nice" or "stupidly infested by senseless icons and page-centered stock images") site promises.
The real number you should be targeting is the number of users you need to pay in order to build a sustainable business, that could be above 80%, it could be below 5%, it could be 0 . . .
This is something that varies widely by business and I don't think it's good advice for a company that falls outside the scope I imagine the author is thinking of.
Consistent user engagement with the product is a better metric to build such a heuristic, because user engagement is direct correlated to value IMO. That's why I am a fan of even charging $1 per month for a product, because it easier (relative to having given it away for free) to increase pricing based on the value you deliver to your customer, which can be measured through user engagement with the product.
It is not. At RethinkDB we had great user engagement, but people were unwilling to pay (or unwilling to pay enough to keep the business going). The heuristic of getting money in early avoids markets with bad fundamentals (developer tools is one instance of such markets).
The concern with new founders working on enterprise products is that it's very easy to convince yourself you're building for the enterprise when you really aren't. RethinkDB is case in point -- it's not the type of product you sell to a CIO (even with MongoDB, very few if any large organizations run on it -- it's mostly limited to silos). We found that we could capture at most $100k in value annually, and $1m contracts were ten years away.
My advice would be to either (a) not start enterprise companies unless you really know what you're doing, (b) start enterprise companies with consumer adoption models (Slack), or (c) if you must start an enterprise company, get really spectacular advisors who regularly spend enough time with you to understand your business intimately, and then listen to them.
That eventually leads to product/market fit (and therefore revenue), but putting a conversion goal on user research sounds more like a sales activity than true user research to me...
Did I just miss the point of the post?
Years ago I signed up for Todoist and I was so impressed by the emails and communication that I got (on top of it being a fantastic product) that I subscribed to premium and have been ever since.
 typical case is, some customer of ours has a weird piece of third-party enterprise software that is flubbing up one of our products. Anti-virus and web proxies tend to be particularly over-aggressive, both with dorking up other software, and with the sales reps.
Some companies which have particularly impressed me with their onboarding communications are Survata (a YC company, https://www.survata.com/ ), and Octopus Deploy ( https://octopus.com/ )