Customer Success

Don’t Fire Your “Worst” Customers. Hire Them a Psychiatrist Instead.'

Jason Lemkin

Screen Shot 2014-03-21 at 9.31.11 AMThere’s a lot of Internet advice that makes a lot of sense, but not necessarily in SaaS.  Somewhat bad advice that entrepreneurs share again and again, without having really tested the ideas behind that advice.

One bit of advice you hear again and again, that I 93% disagree with, is “Fire The Customers That Don’t Fit.  Fire Your “Bad” Customers.”  The ones that suck up too many scarce resources relative to the revenue they generate.  I really think, at least in SaaS, this is some of the worst advice you can get.

Everyone gets the issue(s) that’s been through it.  You get to a certain point, say 1,000 customers, or $2m in ARR, or wherever, and your customer base starts to segment.  One piece of it — say 10% — constantly complains.  And/or, one piece of it — say 10% — uses an edition of the product that you don’t really want to keep investing in and supporting, because you want to invest in the other 90%.  Or more simply, sometimes the bottom 10% of your customer base in terms of ACV just takes way too much time.  The small guys suck up all your customer success and support resources, even though they make up only a small percent of your revenue.

I totally get it.  At EchoSign, the single largest consumer of our customer support resources was our Free users.  The second largest consumer of support resources was our $15/month customers.  The guys that paid > $100,000 a year?  While they had a lot of needs from a solution architecting and onboard perspective … but … from a day-to-day basis, on a relative dollar basis … the least demanding.  They paid 100x or more per year but took basically the same or less amount of customer support resources, and even when you add in customer success, they still took far less resources to support per $ of revenue than the smallest customers.

So fire the small ones, the annoying ones, the non-core ones, you say?

I say think about it differently:

  • First, you will be shocked — shocked — at how little complaints are correlated to renewal rates and customer “happiness”.  The customers that are really going to churn because they are so frustrated they aren’t going to renew … they don’t complain.  They don’t whine.  They just quietly leave the weeks before the renewal is up — for your competitor.  Seemingly out of nowhere, sometimes.  Complaints by contrast are often a sign of deep engagement with your product, coupled with frustration.  No product is perfect.  But these complaints don’t mean the customer isn’t all-in and committed.  And complaints can also sometimes just come with the territory.  Some customers just plain like to complain.  I can tell you from personal experience, the most complain-y customers also renewed at very high rates.  You should hear the complaints, try to address them, and recognize the learning in them.  But don’t assume the customer is lost, or not worth investing time in.
  • Second, those customers you fire may turn out to be the growth layer you want just a few months down the road.  The ones that push your product in a way you don’t want to support today?  That want stuff on your feature roadmap from 2 years from now?  The ones that constantly threaten to leave, and that suck up tons of time from your support team?  You may be surprised that they are actually showing you the future.  A good future, at least a small part of it.
  • Third, even the customers that cost more to support are probably ROI positive if you add in second-order revenue. I know you may think a customer that only pays you say $150 a year and creates 20 tickets a month can’t possibly be ROI positive.  But … what if you keep that customer for 4 years?  And she refers you 2 more customers, one way or another?  All of a sudden, that customer is worth $1800.  Is that really not enough to pay someone $15-$20 an hour to support?  At least while you’re growing, I’d argue it might be.  Not all your customer segments have to be profitable, as long as the (vast) majority of them are.  You can still learn from the ones you just barely break even on.

So what’s my uber-advice here?  If you have a customer segment that is (x) annoying, that you don’t really want today, that sucks up all your customer support time … but (y) that is generating at least $20-$100k in revenue per year as a segment … then don’t fire those customers.  Instead, (z) hire them a psychiatrist.

Screen Shot 2014-10-12 at 5.23.16 PMI mean this almost literally.  Hire a new person to your team, brand new, without baggage.  There’s enough revenue here in the segment, or will be, that you can at least cover her basic costs, this new hire.  He or she probably can’t solve any of the complaints, the problems, the issues.  But she can listen.  Take over the complainers.

That alone can relieve all the stress on the existing team of having to deal with these issues.

This new hire or hires will just listen.  Do on-site visits, maybe, where practical.  Talk through the concerns.  Track the issues and concerns with the customers … but probably not really change anything.  Not burden the product and support teams.  In fact, just be a buffer.

Then, you can keep the revenue.  Capture any positive second-order effects.  And, most importantly, see where these customers take you later.

The last thing you’d want to have to do is re-earn those customers.  It’s so hard and expensive to acquire customers in SaaS.  Keep them as your laboratory to learn from, if nothing else.

Published on October 12, 2014
  • Jason,

    I’ve ready pretty much all of the posts and this is the ONLY one that’s ever left me head scratching. I really want to understand this. Can I ask some deeper questions here, since this is some real pain that I (and probably a lot of people) am feeling RIGHT now, in a real way.

    First, the second order revenue point assumes that the “annoying” segment would actually reference us. Perhaps this is a function of us only hearing negative feedback from this segment, but I find this highly unlikely. Quite the opposite, I find this segment to be notoriously unreasonable in their expectations of us as a software vendor. While I agree that they will likely never leave, I think this just exacerbates the problem; I feel like we’ve got a small base of customers who will never leave us, and always drain us. Never love us, appreciate us, or talk well about us. No matter what we do, no good deed will go unpunished. Hence the appropriately colorful metaphor, “firing” them. What am I missing?

    Second, these tend to be the smaller customers for whom we’ve determined our value proposition is out of whack. Their wants tend to center around features outside of our core product, as they are often unwilling/unable to pay for best of breed software. Far from showing us the future, I fear these customers would force us to build things we really, really shouldn’t (read: a built-in CMS?). What are your thoughts?

    Third, what if you’ve already figured this segment represents an unprofitable customer base? How does that change your point about them being our growth layer?

    Fourth – I don’t understand what hiring a dedicated person would do. Wouldn’t the customers just eventually drain that person as well? And aren’t I adding a FTE (hence more expense)? I mean, if we’re assuming these customers will never have enough, and always be happy, where is the ROI on an additional hire?

    Love to hear your thoughts. One thing I had to chuckle about when I read though, is the ones that leave you never complain. So true. They just stop logging in. We learned that one the hard way, during our first renewal cycle. We quickly put in a customer success platform for exactly this reason.

    • Couple of typos upon re-reading; first, I’ve “read” all of your posts, not “ready”. And on the fourth point, we’re assuming customers will “always be UNhappy”, not happy.

  • Jason, love your work and I think you’re right on here. I wrote something similar with a different type of segmentation here, if you’re interested:

    Sometimes bad customers really are bad, but the ones who *seem* the worst need the most attention.

  • Pingback: Handling SaaS Customer Complaints - Stunning Blog()

Share This