The Old Joel on Software Forum: Part 3 (of 5)

Shrinkwrap support ratios

What's your experience of the number of support calls/incidents generated for each application sold (or bought...) ? I'm talking about shrinkwrapped Windows products that retail for a few hundred dollars.

I was surprised to read that the industry average was about 20-30%. Surprised because I must have bought a hundred or so applications in this category, but have made way less than 30 support calls. At a guess I'd put the figure at 5 or less.

Maybe software is just getting better and better....

Thoughts ?

Count Almasy
Thursday, December 4, 2003

Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning.
-- Cook, Rich

Tapiwa
Thursday, December 4, 2003

You're an atypical user. Most people don't have the background knowledge, won't read manuals, don't google for answers, won't ask questions online. I'm not saying that as a slur, it's simply reality. So when something goes wrong, they naturally turn to tech support.

I can't say I blame them. When something weird happens with my car, I rarely put much effort into diagnosing and fixing it. I take it to the dealer and pick up a loaner car so I can get back to doing my real job while they take care of it. It's not that I /couldn't/ do it, I just don't have the tools or the time or particularly the inclination (even though I like cars).

John C.
Thursday, December 4, 2003

John C

OK, I'd agree with that. But...1 in 3 ?

To add a bit more context: I'm thinking of selling a shrinkwrap app via internet download only. Functionality (and interface complexity) will be roughly comparable to, say, Musicmatch or any typical jukebox.

Let's assume I do a good job with the installer, usability testing, manuals, FAQ, quick start guides, wizards, quality etc etc ... i.e. it's a good product. And by its nature (it's not actually a jukebox) users will be more technically aware than average.

Is it people's experience that 1 in 3 users will want or need customised help ? My best guess was 5% to 10%.

Count Almasy
Thursday, December 4, 2003

It will always be higher than anticipated. My Tech Support staff always seems to grow much faster than my customer base, but everyone I hire is always needed. For the most part, with savvy users, it's not that they need help, but rather, it seems they think you need help. So you will be constantly bombarded with suggestions and feature request.

For proof of that, go to the FogBugz or CityDesk message boards. It seems most comments on those boards are asking for new features with the occasional request for help.

It seems, over time, when customers start to trust that you can deliver what the need and want, then you get even more calls/emails. I long for the days when I could manage all incoming calls/emails because they were from new customers who had easy questions. Now it seems all questions that do come in, take forever to research and fix.

Tony Fontenot
Thursday, December 4, 2003

I agree with the 30% observations. For my main product, about 50% of customers generate customer service incidents. This includes people sending a nice note about how much they like the program, which in a good customer support system, requires sending them a personal thank you note for their support and adding them to a list where we unexpectedly send out token holiday gifts to select preferred customers. Needless to say, the product's reputation is fantastic.

The product is a niche one and supports interaction with many hardware devices. Most support requests are situations where they need help dealing with their hardware object, or want support for a new hardware device added, which we try to do within a week, another 'small touch' that creates a base of customers fanatical about the product. Dealing with personal help for hardware usually takes 1-4 hrs of time. Adding new hardware support takes 10-20 hrs of time, but is amortized across many customers with the hardware and always generates many new sales.

In five years of sales, only two customer support issues were due to a bug.

There are three freeware opensource products that work just as good as our product, but don't come with polite personalized support. Our product costs $999 and is the market leader.

The 33% support issue on most software is the one that makes you realize how commercial software can compete against free and open source software - you have to provide great service.

The Customer is #1
Thursday, December 4, 2003

So, as someone who's often dreamed about running his own software product company, I'm really curious about these statistics.

Over what time period is the 33-50 percent figure valid? One year? One release? Total customer lifetime?

I can't say I'm entirely surprised. This was one of the arguments I recently used to convince my employers that they didn't REALLY want to be a product company. Better to stick with the research and consulting -- and do the occasional custom software project -- than hire a testing team and a support team.

Does anyone have an article I can use to back up these numbers if I need to return to this argument again later?

Jeff
Thursday, December 4, 2003

Have them look into a company called "Microsoft". If ever there were proof that selling shrinkwrapped software is sure financial disaster, they are it.

Mr. Fancypants
Thursday, December 4, 2003

Jeff,

For me it's per customer and I am wrapping all support interaction into one # -- 50% of customers ask for one or more incidences of support, 50% I never hear from again.

In practice, almost all customers self-regulate their requests. We deal with them once or maybe twice the first 6 months, and then never again, except that we'll get thank you notes from some after we do an update with new features.

Only about 5% need extensive support that is frustrating or takes a lot of time.

The Customer is #1
Thursday, December 4, 2003

I've never delivered a product like the one being described, so I have no real numbers or even seat-of-the-pants guesses to offer. But if I were writing the business plan, I'd want to err on the side of pessimism, I think. I have a suspicion that customer support is like time required to complete a feature -- it always seems to come out worse than you expected, not better.

Thought experiment: Imagine we make a bet. You pick a support incidents-per-sale ratio that sounds reasonable and then we look at the actual numbers in six months. I will personally write you a check for $50,000 if you get a lower proportion than you predicted. Conversely, you will give me $50,000 if you guessed too low.

Would you be comfortable making that bet if the threshold between winning and losing is somewhere between 5 and 10 percent?

In any case, I hope you'll let us know what your actual experiences are after you ship...

John C.
Thursday, December 4, 2003

Thanks to everyone for their comments. It seems that I was being optimistic.

John C - point well made. I'll pass on the bet, but certainly report back so everyone can say 'I told you so' :-)

The 30% number came from Soft*Letter (a subscription newsletter), but I've also seen similar elsewhere. The ratio increases dramatically with product cost - high end can be up to 5 per user.

Another perspective, which supports the general consensus, is that every history of a software company I've ever read always includes a tale of Customer Support being overwhelmed and needing urgent upgrading.

Count Almasy
Friday, December 5, 2003