Better cancelation flows
kylewritescode.com
kylewritescode.com
Step 1 (after clicking Cancel My Account): "We often find that people cancel for one of the following three reasons. Here's our most persuasive microcopy possible on why that wouldn't be a good idea, two sentences each, with links to longer explanations. [Wow, That's Awesome] [No Thanks, Cancel My Account]
Step 2 (the genius bit): "OK, your account is canceled. Your service will remain active until $DATE. We will happily extend this to $DATE + 7 days and then cancel your account with absolutely no funny business if you take this brief exit survey."
Was the account really cancelled at the end of the first sentence? It sounds like it won't actually be cancelled until either $DATE or $DATE + 7 days.
And you don't want that, because they lead to chargebacks.
This seems a lot like the situation where an employee threatens to leave and the company offers more money to keep them for which I almost always give the advice: "just say no".
I guess if price is really the only thing making someone leave, then lowering the price might make a difference but price and value are intrinsically linked: if they thought the price was too high, then they considered the value to be too low and it's probably just a matter of time before they decide to leave anyway.
Lots of people think that lots of business will treat them better if they are "loyal" and stay with a company for a long time. If you're a bog standard residental customer and it's a large company with thousands or millions of customers, this doesn't hold. Usually those customers get some of the worst deals.
Forget loyalty. This isn't a marriage or an army, this is business and economics.
Not to mention, shouldn't every Fancy Hands customer that reads this go in and pretend to cancel to get the discount?
This is a falsehood believed by engineers for nebulous reasons largely founded in naivety about business. Businesses routinely offer the same service at multiple price points. Virtually every input in your business is offered at multiple price points, from paper to telephone service to your Internet connection to your VPSes to... you get the general drift. Variable pricing is an observable and unremarkable fact of life. It causes virtually no drama.
Not to mention, shouldn't every Fancy Hands customer that reads this go in and pretend to cancel to get the discount?
Want to get a free shake from McDonalds? Buy a shake. Consume it. Take a napkin from the napkins, put it in the shake cup, bring it up to the counter, and say "I am dissatisfied with my shake, because it has napkins in it. I want a new shake." You will get a new shake. Why does every customer at McDonalds ever not scam McDonalds for free everything? Partially because they have principles, partially because they fear social opprobrium, partially because they are unaware of the opportunity, partially because they just can't even conceive of why any sane person would scam McDonalds for free shakes, but it is an observable fact that McDonalds collects revenue for nearly all shakes despite there being (numerous) pathways to stealing them.
As I discovered last week, even if you throw your half-full drink on the floor thanks to shaky-handed incompetence, walking up to the counter apologising profusely and asking for a mop will achieve the same result. Along someone else for you to apologise to when they arrive to mop up your mess.
Ah, the credit card company gambit. Then you can charge more for default accounts, to allow the subset that fail to see the game being played to pay for the those that do know the trick.
Won't most of your other users catch wind of this and fake cancel just to get the discount? And for those that don't or don't feel good about faking a cancel, you're essentially charging your happy or loyal users more. Backwards from a business practice perspective and wrong from an ethical one.
Ask a simple question, get a simple answer: no.
More elaboration:
This misconception comes from a few places.
Many HNers passionately care about money. Many customers don't. In particular, the best customers (largely businesses) perceive not even the smallest scintilla of pain from software bills ranging from $20 to $20k per month and the person who'd actually be pushing that button at the business perceives no difference in their personal lives based on what the price is.
Additionally, we live in our software, are familiar with all the intricacies, and assume that they matter. Our customers deal with our software only sporadically, and generally only allow it to distract them from things that matter to the extent it is required to accomplish their goal. They don't typically go spelunking for hidden options. They (largely) don't go to CustomersOfFooApp.com to swap hidden secrets about FooApp because customers of FooApp care about FooApp about 1/10000000th as much as the developers of FooApp do.
Finally, most people have what you might describe as either a moral compass or a fear of embarrassment which will counsel them that attempting to scam businesses for discounts is not something they want to be doing. (This assumes you're not dealing with a customer population which is toxic by nature. Charge more.)
Some gamers speculate on the frequency at which the service allows this.
I've used this to my client's advantage by seeding coupon sites with special coupons. They don't get much of a better deal than anyone who sees other promotions on the site, but you can trick them into spending more money with minimums.
On the second half, it's a fair point. It's too early to say, but certainly worth keeping an eye on.
If you've never been through it, here's what it boils down to:
- "Cancel my account" link was hidden and relatively hard to find. (I guess this one's normal)
- You have to give a reason why you're quitting, out of a major radio-button list that felt more like a survey.
- No matter which choice you pick, you get some bogus argumentation on why you shouldn't close your account. I bet some manager in the company is so happy about how much this would reduce the number of possible cancelations. I'm sure she even has some bogus data to back her claims up. To me, it felt like an uber annoying phone sales representative who won't let you hang up the phone even though you're clearly not interested in the product.
- If you pick the field 'Other', you have to fill the text field explaining why.
- After all this, you have to enter your password and/or read a captcha (I don't remember exactly)
- It ends with Facebook telling you that your data is still there, ready and waiting for when you'll be back. I don't remember the exact wording, but to me, it came off as extremely smug.
My point is, canceling has to be easy. It has to be so seamless I could do it in a few seconds. It's a matter of respect. I felt that Facebook didn't care about making me hop through their hoops. Their final message, that I'll be back, added to the insult.
It's a great thing to survey cancelling customers, and yes, you could probably learn a lot and mayve save a few of them. But for the love of God, let your surveys be as non-intrusive as can be.
Either way I handled this all through email, which made it easier to fully understand their problems, and in many cases turn it around (since every month is different, it may just take one cycle to get them happy again).
Most of the people canceling relayed that they were still passionate about what they were receiving, so what I've been doing recently is giving them free subscriptions and having them on as brand ambassadors, helping me spread the word about Startup Threads Monthly. Too early to tell if this is a good idea but its what I'm working on now.
That's one thing I'm also interested in exploring: What can we do to buy us one more month to help them find real value in the service.
> Now, when a user wants to cancel, we give them two buttons front and center... If it’s because of price, we’ll give them an instant, lifetime discount;
I might need a LOT of help for 3 months, then no help for 2 months and then again for another 3 months.
I had to cancel my LegalZoom account because I didn't need it for awhile, but then opted to not sign up again even when I did need it again a few months down the road.
If one subscribes to linguistic descriptionism, which means "The right way to spell it is the way that people spell it and expect to read it", people do indeed frequently spell it with one L.
Firefox shows "cancelation" as a typo in both their UK and US dictionaries, although it appears in /usr/share/dict for both american-english and british-english on Ubuntu. Curious. I would have indeed natively spelled it with two L's.