Hacker News Reader

Comment by danjac | Hacker News Reader

It's particularly common with big whale enterprise clients, where the salesperson is trying to close a very long sales process or the client has you by the short hairs in the contract. I remember one case where we had to implement a very complex and time-consuming custom SAML integration to a platform where we just had one client asking for it, because that client was too big to ignore. Once we integrated SAML, the number of end users actually using the SAML login option was in the single digits. It was just a checklist item for some purchasing officer flunky.

View on HN

For SAML in particular I can definitely relate here. I will say though, that for SaaS products that you want large companies to use, you owe it to yourself to have an auth layer that is immensely flexible and pluggable. Auth is a huge mess and large companies nearly all have single-sign-on solutions, sometimes strange bespoke ones, and so you gotta expect to potentially build a zillion different auth integrations if you want to sell to large companies.

This is exactly why we built WorkOS. Check it out if you’re looking for SAML/SCIM for your app. (I’m the founder.)

https://workos.com/single-sign-on

Sure, but if you are planning to go from SME to enterprise you need to think strategically for the long term. You'll probably need to have more developers to deal with the complexities and hand-holding you've touched on (auth is just one of the issues and not even the most complex) and probably non-technical roles for all the compliance and so on.

The danger of landing just one or two "whales" is that you bend your product backwards for these clients, but without thinking through the TCO it can cost you more in the long run and put your business in more jeopardy than when you rely on a few hundred SMEs.

The fact that no one used it doesn't mean that building it was a mistake. Maybe the purchase officer got it wrong, but the salesperson who requested it be built got it right, since the deal did in fact depends on it.

Perhaps. But then it's important to look at the total cost of ownership: was it worth building a complex feature for one client that adds to the technical debt of a code base and tied up developer resources that could have been better employed elsewhere? Could we have pushed back on this feature or was it really a deal-breaker?

In other words, perhaps landing this one client was important to that salesperson, and perhaps in the short term to the company, but in the long term it may have been better to have lost that one deal. Very often these decisions are made with little calculus about the TCO.

Reply on news.ycombinator.com