Comment by patio11 | Hacker News Reader

Comment by patio11 | Hacker News Reader

For Appointment Reminder, approximate per-account customer support incidents per month. I've taken the liberty of scaling them to X, where X represents the number for the highest publicly available account plan.

  Personal ($9): 7X
  Professional ($29): 4X
  Small Business ($79): 3X
  Office ($199): X

The character of the questions is also different at the various plan levels. Most common question for Office: "What's the timeframe on integrating this with ..." followed by "Our $TITLE would like a report saying $NEEDS, can you make that happen?" Most common question for Personal: "How do I schedule appointments?" followed by "The system is working exactly the way it says it does on the screen. Can you please tell me why that is happening? I thought it would work in a way completely opposite to the way described on the screen. It would be convenient if you could fix that. No, I didn't read the 'If you want this to work in the opposite fashion...' text on the screen to change that setting, I have more important things to do than worry about computers."

Your mileage may vary. If I were doing the math based on phone calls waking me up in the middle of the night, the numbers get skewed due to one pathological customer in the $29 bucket, who has literally called me more than every other customer combined.

P.S. I have fairly exact privileged information regarding this question at a handful of companies and anecdotal evidence from dozens of my software buddies. It is our universal experience that the support load for cheap/free customers crushes the support load for the higher plans, both on an absolute and per-customer basis.

I noticed a similar pattern in my business (I run an appointment scheduling web site). An additional insight was that when you provide a quick reply to the "dummies" they will start asking even more questions instead of trying to find stuff out themselves.

Here's what works for me: on the second "silly" support request they send I put them in a "cool down" queue that gets answered a day later. If they keep asking simple questions they move into the two day wait queue. (Obviously, this only works if your support is by email.)

I worked as a consultant for a company once where they would go even one step further and would actively encourage customers that used too much customer support to switch to another provider. The trick is doing this in such a way that the customer does not realize he's being told to leave lest you get a bad reputation. I'm not going to divulge all my trade secrets here, but I've found a way to do something similar for my business, perhaps you can think of one for yours too.

I don't think managing expectations is really a trade secret.

A customer losing strategy is opposite to most startup who are looking to gain customers. It's a good problem to have and a sign of headaches that need to be managed.

If you could share a bit more on how you help the customer realize they might want to move, most startups are trying hard enough to acquire customers, not lose them.

I'm not really trying to be secretive but there are several details about my business you would need to know to understand my specific "selective denial" method (as strategy consultants call it). But the general gist is this:

Many startups are so desperate to find customers that they are even holding on to the ones that cost too much support. (Note that in a startup support cost often takes away valuable time from the founders and thus is more an "opportunity" cost than a monetary cost, they should have spent that time growing the business with more valuable customers).

I chose to do only e-mail support because it self-selects out a type of customer that wants to use the phone (also phone support scales poorly, especially when expanding internationally). There are ways you can encourage the "dummy" user to desire phone support, and then gently point out that there are _other_ providers that have that. If you do this properly the dummy client will actually feel they've had good customer service (and you could argue that they have, we just weren't meant for each other) and your over eager competition gets stuck with the dummies.

Edit: I thought I should mention, for a startup it can often be unclear which customer is/will be valuable and you can learn a lot from the "wrong" ones, so be careful with "selective denial" unless you've run the numbers and are sure what you are doing

In writing a comment on Sacha's blog, I came to the idea of deciding to offer a simple response to support incidents coming from the lowest paying tier. I wonder if anyone has done this and if you think it would work, or be a smart move?

Create a canned automatic response to someone submitting a request/support email that goes something like this.

"We have worked really hard to make this app/service as easy and intuitive to use as possible. On top of that we decided to price it very low to make it even easier and affordable for people to take advantage of. 80% of our paying customers never contact us for help or support. We honestly just don't have the time or resources available to explain how to use our service. If our help section and FAQs can't fix your problem, then you probably shouldn't be using this app, or we're gonna have to ask you to pay for support.

Option 1: Click here and we will immediately cancel your account and refund you.

Option 2: Click here to upgrade your account for support

Option 3: Figure it out!

Have a Nice Day!"

Or something like that. Perhaps it can written to be more friendly or in a more suitable tone. But you get the idea.

That's a rude and unprofessional answer.

If you don't have the resources for providing support and can't help yourself being rude, then maybe you shouldn't give the free/cheap option in the first place.

There are other options besides behaving like a dick.

Personally I like how mobile operators handle this. Whenever I call their free support line, I am placed on wait for at least a half an hour, with a message along the lines of "our operators are currently busy, please hold ; if you'd like priority then call this number ...".

If he doesn't have the resources for providing support and can't help from being rude, maybe he shouldn't be providing subscription-based software products at all.

Software is support. Sure, there are ways to mitigate the cost and quantity of support, but once the product is built support is one of the primary functions of the business.

The wording is rude, but I'm sure that's because it's an off-the-cuff answer. The concept is valid though - advertise up-front that the lowest cost option does not include free support. You can add a 4th option with a link where users can ask for help from other users.

Thanks for providing the rough figures, it's quite rare to see numbers on things like support requests broken down to pricing tier but quite fascinating. Do you have any insights or intuitions as to the varying number of support requests?

Is it that the Personal plans are done by small businesses where the owner is trying to run their shop and do IT work whilst larger businesses tend to have a knowledgeable sole-purpose "tech guy"?

Edit: Ah, I hadn't seen that you've written so much about it previously. Thanks! For others, this seems like a good overview of low tier pathological customers by patio11 -- http://news.ycombinator.com/item?id=3186111

(for more see patio11's suggested hnsearch.com [patio11 pathological customers] at http://www.hnsearch.com/search#request/all&q=+patio11+pa...)

hnsearch.com [patio11 pathological customers] will cover that in a lot of detail.

What is the ratio of customer types itself? Reading, this, I wonder

1-If the are just 7x more people in the Personal bucket. (I'm guessing you account for this)

2-If there are so many more in the Personal bucket that it is still the primary driver of your revenue, so that is why you keep it?

The numbers are per-account, not summed up.

But I bet you have many more $9 and $29 than $199 customers?

I'm sure the equation still favors your theory, but I'm also sure by a lower number than what you'd like readers to believe.

For Appointment Reminder, approximate per-account customer support incidents per month. I've taken the liberty of scaling them to X, where X represents the number for the highest publicly available account plan.

you are certainly right with your scale, but i guess zobzu's point is, that your non-support costs (say: development) is mostly paid by your 9$ and 29$ customers (in absolute terms).

so you cannot just get rid of every customer but your high paying ones.

Good point. Of course it's possible that 7X is a cost that cancels out the revenue from those customers, in which case it would be better not to offer the option, but as long as it's not, it's still profitable to offer the plan. Depending on the ratios, the cheap customer bracket could even be more profitable (in absolute terms, as you say). You just have to build the extra support cost into your pricing.

$9 customer - $5 support = $4 per many customers

$199 customer - $1 support = $198 per few customers

Reply on news.ycombinator.com