This is an incredibly important lesson, and one which is counterintuitive for a lot of people in SaaS, because you *almost never* hear from your best users.
If you e.g. figure out product priorities by listening to what your last 100 support tickets, they will be disproportionately from low-value plans with users who are fundamentally dissatisfied with offering and will not, with an incremental improvement in the product, be happy.
"Aren't your most valuable users the hardest ones to support?"
The exact opposite of the case. Normalized by account it's (often) the little users, by lots. Normalized by dollar value it isn't even funny.
The exact opposite of the case. Normalized by account it's (often) the little users, by lots. Normalized by dollar value it isn't even funny.
Why? Many reasons.
1) Because they have internal teams to support them on functions that otherwise get kicked to you. If you sell a $19/mo SaaS you have to bundle in a lot of basic IT support; if you sell $999/mo SaaS then they have an IT department for that.
1) Because they have internal teams to support them on functions that otherwise get kicked to you. If you sell a $19/mo SaaS you have to bundle in a lot of basic IT support; if you sell $999/mo SaaS then they have an IT department for that.
2) Because you're getting better, more sophisticated users towards the top end of the scale. They know the problem domain. They make less mistakes that they need to open a ticket to rectify. They comport the rest of their business in sophisticated fashion, too.
This is the "hard drive failures are randomly distributed except that they're observationally not" thing that hits every SaaS company at scale.
You will have tickets whose underlying cause is "my hard drive failed." They will disproportionately be from your lowest-value accounts
You will have tickets whose underlying cause is "my hard drive failed." They will disproportionately be from your lowest-value accounts
Your engineer brain says "How does your commodity hard drive understand that you don't pay enough money for SaaS and decide to brick itself to protest?!"
That's not what you're seeing.
That's not what you're seeing.
You're seeing a very slight difference in failure rates (cheap disks vs production-worthy disks, better care) and a large difference in "Raise your hand if you have no plan in the event of a hard drive failure because you don't operate your business like that is a thing."
(Some people who have not worked CS for a SaaS company are wondering "Why would it even matter if the hard drive failed?"
There are many variants of this: "My hard drive failed and I need a year of invoices for tax purposes." "My hard drive failed and I lost report." etc
There are many variants of this: "My hard drive failed and I need a year of invoices for tax purposes." "My hard drive failed and I lost report." etc
And you think "Huh if people need to download a year of invoices why don't we just make that possible" and then you do. The top end of the LTV distribution immediately finds the feature and starts using it *and the bottom end does not.*
3) The more sophisticated a company is (price a good proxy for that) the more likely they have identified the right tool for the job, and are not attempting to pound in a screw with a hammer.
You wouldn't make a seating chart in bingo card software. Trust me: many people would.
You wouldn't make a seating chart in bingo card software. Trust me: many people would.
(What's the one thing bingo card software does? Randomly distribute N elements over each printing of 5x5 grid. What's the one thing that a seating chart cannot do? Randomly distribute elements over each printing of the grid.
I literally built a Seating Chart Mode. Not a joke.)
I literally built a Seating Chart Mode. Not a joke.)
