Case studies on how to drive away your customers

If you ever wanted to lose business, you'll love this!

Posted on May 11, 2023

Are you a business owner or executive with too much money? Do you want to just chill, but customers get in the way by always complaining? Here are some best practices from industry leaders on how to tell off your customers!

Disclaimer: I'll make fun of some companies in this letter. This doesn't mean that I think they're altogether terrible, or that I wish them ill. In fact, I like most of these and their services very much. However, there are cases when they messed up, and those are teachable moments we can all learn from, not to commit the same mistakes.

Atlassian would like you to just please go to their competitor

If I had a nickel for every time Atlassian discontinued a service, made me trust them less, and forced my company to buy the same thing from Microsoft then I'd have 20 cents. Which is not much, but it's weird it happened four times. Almost five, actually.

Atlassian is notorious for rewarding you being heavily invested in their ecosystem with breaking your stuff and then telling you to just deal with it. We at Lombiq have been Atlassian customers for a decade, and during this, I lived through them discontinuing the following:

  • The cloud version of Bamboo, their CI system. At the peak of us using Atlassian services, when we believed we perhaps wanted to do everything with their products, we were in the process of migrating over to Bamboo from TeamCity that we self-hosted. Then Atlassian made Bamboo self-hosted only. Right at the time when CI SaaS options became mainstream... We use GitHub Actions now.
  • HipChat: It was a great simple chat service that started to show its age though, due to the lack of attention from Atlassian. At one point they announced Stride, their Slack competitor and slowly nudging HipChat users to move to the yet-beta Stride; we almost started migrating over. Then as a spectacular plot twist, they shut down both HipChat and Stride, urging people to use Slack instead, which Atlassian acquired a stake in. We said peace oot and moved to Microsoft Teams. To give credit to Atlassian, they did offer an automatic migration to Slack, and a full export feature if you wanted to move elsewhere.
  • Confluence Team Calendars: Confluence, the wiki engine of Atlassian had a simple calendar feature. We used it to display project-related dates and have a common calendar for e.g. team events and holiday leaves. This was too much for Atlassian, so they moved it to the Confluence Cloud Premium offering. Since that contained only things we otherwise didn't need, instead of paying $5 per user per month for a simple calendar, we moved to Outlook which already came with Microsoft 365 we otherwise used.
  • Mercurial on Bitbucket: Mercurial is a great source control system that was once a contender of Git but in the end, Git won. One of the last and certainly the biggest holdouts of a hosting service for Mercurial was Bitbucket by Atlassian. Right until they closed it down, keeping only Git hosting, and did little to help people cope with their repositories getting deleted at a cutoff date. Perhaps spending some time on offering an easy way out with a "Convert to Git" button (a feature that there are actually open-source projects for) would've prevented people from showing up with torches. Moving all of our private repos to GitHub, where we already had all of our open-source, was a no-brainer.

Things that would've helped me stay with Atlassian:

  • Not leading me on by enticing me to use their service just before kicking me in the shin by closing it down. I.e., communicate clearly and openly with your users.
  • Offering some help with a migration to stay with them. But if I have to do all the work, and with the same work I can go to a better service (Outlook, GitHub), then I'll go to a better service.

Also, offering an interlinked bundle of services works as a business strategy, but even this won't save you from people moving away if you piss them off enough. Especially developers, who can code their way out.

At GitHub we value your contribution to help our users, but not enough to not delete it

I'm all for open-source and open knowledge. Thus, if I spend my limited amount of keystrokes to help someone, I'd like that to be available to as many people as possible. I.e., if you have a question, I'm happy to help, but out in the open. And when people help me, it's a magical thing, and I'm really grateful for it. But then GitHub comes and deletes it.

GitHub had a central Discourse discussion board where people could get help from each other for everything involving GitHub. You know, increasing the value of GitHub.

At one point, they decided to move to their own discussion board, which is fine. However, in the process of this they a) made the old board unavailable and b) only moved those discussions over that had a message marked as a solution. If a thread had a page of valuable information but no marked solution (perhaps because there was no one solution, just workarounds)? Well, try to fish it out from the Wayback Machine or Google cache, and have fun with that. I had two important discussions vanish that apparently didn't get the memo that when something's on the internet it's there forever.

The takeaway here is, don't destroy knowledge that your users share with your users, for free, so both of them can pay for your service more happily.

Basecamp is happy to hear your feedback if you'd please repeat it every six weeks

The project collaboration tool Basecamp is usually great. When it's not great, you don't need to bother giving them feedback though, because if it doesn't pique somebody's interest immediately in the next development cycle, it's thrown out (huge kudos for them to publish this book for free though). I think I get the no-backlog concept. However, as a user, it demolishes my motivation to provide any product feedback if how it'll be taken into account is an opaque process, and with a high chance, it'll be thrown out without much thought. I had feature requests that I submitted to Basecamp support over the years, and I guess all of them ended up in the void.

Atlassian has taken a lot of heat from me above, and here I'd like to commend them for maintaining public issue trackers for their products. (Now, don't ask them though if a top-voted feature request will wait a decade until it's actually implemented.) As a user, I favor this openness, the ability to cooperate with other users, and that my keystrokes don't go to die on Basecamp's betting table.

Another hugely demotivating way to manage user feedback is similar to the GitHub discussion board fiasco above: I've seen both Microsoft and Zendesk collect feedback in a UserVoice idea board. They then deleted these, without bringing the previously submitted ideas over to the new system. I really didn't have the willpower to resubmit my feedback.

No, I still won't give you my phone number

This is not unique to one company, but I see it frequently. Your company has some stuff that I need and would like to buy or have a question about. Awesome, we'll make so much money together! Now if I could just write you an e-mail or something... No, there's a contact form with name, e-mail, phone number, and a kind message that you'll call me back! ... What? Can I at least attempt to tell you my very specific question before we chase each other via phone?

A sub-species of this is when I'm able to send them an e-mail. Then instead of getting an actual answer, they request a call, without reflecting on anything in my e-mail. Please, no. I had one case where we scheduled the call but then the sales rep didn't show up. Stellar. Alexa, play the hymn of salespeople.

Please, let me write you a message. If you need more info after that, or your solution is really tailor-made, then we can chat. But we won't start with a call from zero. Let's respect each other's time a bit more.

Moral of the story

I guess, follow Wheaton's Law with your customers?


Illustration: Midjourney prompt "Annoyed person crying about HipChat being discontinued by Atlassian. --aspect 3:1 --seed 3516117457"