I tend to disagree with 37Signals on a mess of things. Like a lot of successful internet pundits, they deal in absolutes and hyperbole. There’s no middle ground and there’s no “…well, it depends”. That’s just not as linkbaity. It’s probably not as fun, either.

But there’s one place where I wholeheartedly agree with ‘em– if you’re in the Freemium game, start charging for your software. Right now. Yesterday, in fact. Should you put a price tag on just any web service? Absolutely not. Kayak shouldn’t charge to find you a flight and (if the rumors about their success as a leadgen platform are true) Mint shouldn’t charge you to organize your personal finances. But if a big part of your revenue plan involves charging for premium services on top of a free product (freemium), you should start charging as soon as possible. Here’s why:

1. Price signals value. Where you set your price emotionally sets a value for your product. What that means is that amassing a gigantic pile of enthusiastic free users isn’t going to result in a big pile of paying users when you turn on your premium features (or worse, move some of the features behind the “pay wall”). In fact, it will likely piss off a lot of users who have grown accostomed to getting something for nothing. During the Y Combinator experience, I got a chance to hang out with Joe Kraus (founder of Excite and later Jotspot) and the fellas from Wufoo. Both had horrifying anecdotes about asking a bunch of free beta users to start paying for their software. The conversion rate was awful. When we first turned on our premium offering, we were struck by the same thing. We opened the floodgates for paying customers and found that almost none of our free users made the switch. So even with your pile of zealous free users, you’re starting from square 1 in the premium game– you’ve already convinced your current userbase that the fair price for your product is “free”.

   I mentioned this in the comments but I wanted to promote it up here as well. *”Take a minute and answer this two-part question:*

   “1. Is the percentage of African nations in the United Nations higher or lower than 65? 2. What is the percentage of African nations in the United Nations?

   This was one of the queries that Amos Tversky and Daniel Kahneman posed in their 1974 paper in Science called “Judgment Under Uncertainty: Heuristics and Biases.” It turns out that the answer you provide to the second question is heavily swayed by that first question. The average estimate for question two was above 45 percent. When question one was lowered from 65 percent to 10 percent, the average estimation of question two was dropped to 25 percent. ” Source (pdf)

   Your free beta anchors your perceived value at zero and it’s a bitch to climb out of that hole.

2. Speaking of Square one… You don’t know nothin’ about square one! Charging people money for software is a whole new set of skills that you quite likely don’t have. What do your paying customers REALLY want? What do you put behind the pay wall versus in front of it? What do you give your free users? What kind of free trial should you offer? Will a referral program work for your business? Does your value proposition resonate most with individuals or businesses? Big biz or small? Can you make adwords work for your business? What works on an adwords landing page? Will a salesperson be valuable for your business? Where do your leads come in? Telesales? Direct Mail? SEO? SEM? Viral/word-of-mouth? The problem with all of these questions is that the answers don’t transfer across markets very well. What might work great for my market/product might perform terribly for yours. The sooner you start investigating this stuff is the sooner you start being smart about your market.
3. Getting people to sign up for a free service doesn’t mean that you know they’ll spend money on it. There are lots of clever ways you can ascertain whether someone would REALLY buy your product. You can put up fake adwords ads, you can cold call people, you can throw up a permission marketing page and try to get attention for it, you can do a focus group, you can ask some pricing experts, and more. But nothing is a perfect substitute for having a buy button next to a price and seeing if anyone actually clicks on it. And they generally won’t. At first. So start learning!

We’ve been at this for almost two years and I have very few big regrets. But my biggest regret as an entrepreneur is not starting on the path of charging customers sooner. It’s taken us about a year to get pretty good at it, but we’re still learning new stuff about our customers every week (we’re pretty darn grateful to have customers who are generous with feedback).

Some additional fabulous reading on the topic of when to charge can be read on Sean Ellis’ blog here. Sean basically contends that you shouldn’t charge at all until you are certain you have product/market fit. In the comments, someone expressed concern that product/market fit isn’t real until there’s a price attached to it. Here’s Sean’s response:

I agree that price is part of the process of figuring out if you have product/market fit. I’m basically starting with the price of zero. If people aren’t that disappointed to see the product go at zero cost, then we already know that any cost above zero will very likely also result in people not being that disappointed to see the product go. Once enough people consider it a “must have” at zero cost, then the next step is to figure out a price that generates the most revenue for every thousand people that try the product.

This is an interesting thought, but I’m not convinced. I remember hearing that Wufoo and Jotspot both had pretty passionate free/beta users. I could be wrong, but I’d wager that they would’ve had a solid number of folks who would state that they’d be “very disappointed” if they had to give up the product. Nonetheless, they came up pretty empty when asking these users to start paying up. The difference between product/market fit for a free product and product/market fit for a $5 product could be a lot farther than you think. It might be a few iterations or it might be a whole new product.

But where I think Sean is absolutely right (to be fair, I think Sean is brilliant– you should subscribe to his blog!) is that you need enough customers to be able to measure and improve your product. If you can’t acquire/retain 100 paying customers, perhaps you should stick with a free/private beta.