Thursday, March 22, 2007

Lessons Learned #2

The next biggest thing I got out of this product experience is, be careful with market research/pricing.

Originally the plan was to make software that could be sold to all "service industries", you know, those guys in the white vans that do all the dirty stuff we can't do. I mean, they all basically do the same thing? That was the thinking at least. First we would use chimney sweeps as our test market, have them give us all the feedback early before we start dealing with plumbers and electricians.

People talk a lot about market research, but I have yet to really find a good guide on what you should do. I think because a lot of it is just intuition and experience. This is where working on something you know well comes in handy. For example, Ian worked a lot with Help Desks, and Patrick was a teacher. I on the otherhand, have never been a chimney sweep, or any other type of service professional. In fact, I just paid someone to fix some drywall in my house, which could quite possibly be the easiest home improvement project out there, that's how far removed from the service industry I am. But, remember I started this with two partners!

The big partner with me was the guy that had the original idea, and was a chimney sweep. I'll go into Partner lessons learned another day, but for now I'll just say that how I did my market research was based a lot on him. My experience in the retail end of the software industry had been with two companies, one sold software for court reporters (which would seem like an even more obscure niche) and the other sold web apps to enterprises. The court reporting software sold for around $4500. The enterprise software sold for an average of $30,000 a year. So thinking of selling software for $19.95 is frankly depressing to me. Of course there is no good reason for this thinking, because if you sell 2 copies an hour, your making some decent scratch, but I was definitely used to the big sale.

I knew from the start ChimSoft couldn't sell for $19.95. At that rate if every single Chimney Sweep in the world bought a copy, that might pay my mortgage for a month or two. So here is my super secret market research technique:

1) I asked my partner how many chimney sweeps he thought there were in the USA. He said his certified sweep number was 5 thousand something, and he knew they were consecutive and that they started issuing them in the 70s, therefore there must be around 6,000 certified sweeps out there.

2) If there are 6,000 sweeps who received certifications, you have to figure at least half are dead, retired, or gave up, leaving 3,000 active certified sweeps.

3) Most sweeps are not going to get certified. It's expensive, and time consuming. We'll assume 1 out of every 4 sweeps is certified. This gives us 12,000 chimney sweeps in America.

4) Now, some people will take that number there, multiply it by $100 and start shopping for a retirement villa. But this wasn't my first business. I knew at BEST I'd ever sell to 2% of the entire chimney sweep market in a given year. So best case scenario we are talking 240 sweeps a year.

5) Now we pick a price. I started this way. "How much will this need to make a year in gross sales to make it a worthwhile business?" I'll say $300,000. So in the absolute best of times the business should be making $300,000 a year. (Keep in mind that I was not in the mISV community at this time, and also that I have two partners, and no huge desire to quit my job, so a low ball $25k wasn't going to cut it for this).

6) So when you put all that together, assuming you have the best scenario of 240 sales a year, to make $300k, each unit must be $1250

7) Ask the customer if they will pay $1250 for the software

What? $1250 for a stinking piece of software? From chimney sweeps? Well, the thinking wasn't that crazy. We debating the price a lot, but the chimney sweep partner continued to point out that we were not selling to individuals, but in fact businesses that can deduct on their taxes, etc.. Chimney sweeps also have been proven to pay big bucks to help their business as most own $4,000 chimney scanning cameras. Heck, our software can replace most of the functionality of that camera and it's only half of the price.

It's hard to say that this price point was bad. We went to our first convention with a price tag of $2000 for the software. There were a few sales at this price point, and LOTS of interest. We talked at great lengths with people who hemmed and hawed about the price, and continued to keep in contact with us through email, phone calls, etc... Could we have gotten triple the number of sales if we priced at $300? Probably, but we would have ended up with less money. In hindsight, I think I would have rather had more users spreading the word of the program then the actual cash.

So what was the problem with all this thinking? Sounds like it was priced about right (amazingly, considering my "numbers out of my ass" market research method). The problem was that the number of sweeps was VASTLY overestimated, despite my huge cuts. A year or so into the project I was talking to the sweep partner again, and he mentioned something about there being 2,000 certified sweeps. Somehow we had gotten confused or miscommunicated or something, but that original 6,000 number I was working with was actually 3 times the reality.

Now that I've been in the chimney sweep world for a while, I think the true estimate of active sweeps in the country is somewhere around 2,500. Again, that number is basically coming from intuition and estimates. And while they will spend $4,000 on a camera, they do this because they can charge $400 more per sweep to use a camera. And the camera can find $3000 of work to be done. There is a direct link to them between profit made and money spent. Most service industry folks still do not see software this way. Unfortunately most chimney sweep businesses also do not have a lot to spend. I heard a lot of comments at conventions about how the whole industry is suffering, so I had a lot of sweeps asking me for financing plans. I just could not do this, if you've seen my BudgetSimple philosophy, you know I hate debt. So at $2k, the price was too much for a poor chimney sweep, and the big chimney sweeps didn't want to pay $2k X 8 vans + 3 office computers. Once dropped to $300, sales increased, but alas, based on the number of sweeps out there, it's just not enough money to motivate me.

Why not continue the original plan though? Sell to plumbers, etc? Well, ServiceCEO is doing just that, but the program becomes too complicated for most of the end users. I started to do it with ServiceGadget. Unfortunately, you really have to learn and market to each of those industries differently. Not being able to use the product myself, I didn't have a great deal of interest in this prospect after spending 2 years on ChimSoft. I have too many other ideas that I am really excited about.

Also, don't let anyone tell you that as a one man business you can just keep adding products and make more money. Every time you work on a new product, you take your "eye off the ball". As a one man show, you can only physically (and more important mentally!) dedicate to one product at a time. So when I'm working on ChimSoft, my other businesses and ideas languish. That's part of the reason I had to have a clean break here, just to free up some mental focus.

2 Comments:

At 1:46 PM, Anonymous Anonymous said...

Wow! I followed your blog through Planet ISV, and had made a subconcious estimate of your sofware costing ~$49.

I sell to businesses. Not quite such a vertical market, but your post has certainly got me thinking about pricing...

 
At 7:35 PM, Blogger Phil said...

Anon, if you sell exclusively to businesses and your program is more then a small utility, you probably are underselling yourself. Speaking as someone who has been responsible for purchasing at several companies (including my own). As long as software is under $2k or so, it doesn't really matter what it costs. The boss or whoever won't care the difference between $50 and $300 as long as a purchase order isn't needed.

 

Post a Comment

<< Home