Jason Cohen’s analysis of “marketplace” business models is fantastic, but this gem is the most important:

But just because automation is the goal doesn’t mean it’s the way to start. The good thing about automation is it’s efficient; the bad thing is you cannot learn because you’re not involved in the process. And at the start, learning is where you should be spending most of your time!

For example, when SpareFoot began they weren’t sure how to charge storage companies. Should it be a $20/month listing fee? Or a flat “finders fee” per lead? Or a commission on leads which convert to sales? Could they charge extra for a “premium” listing? Should purchases go through SpareFoot so they can extract their cut, or should they bill storage companies separately?

If they had picked a strategy and automated it, there’s a low chance they would have picked the right one. All that time spent writing and debugging code, worthless.

Instead SpareFoot decided to automate nothing. When a potential buyer made a search, they grabbed their email or phone number and said “Thanks, we’re going to find you a great deal by Thursday.” Then they banged the phone all day, calling up regional storage facilities. Their pitch was awesome: “I’ve got a lead for you; his name is John Doe and he’s looking for a 10×20 with air conditioning. If your rate is competitive, we can do the deal today. By the way, if you want us to send leads like this to you all the time, it’s $20/mo to list with us.

The bold phrase in there is damn compelling, right? And of course they varied the offer based on current pricing theory or in real-time based on the interaction with that particular storage facility.

None of this — determining the pricing structure and amount, building relationships with facility managers, and ensuring buyers’ success — would have happened if they started by writing code or any other sort of automation.

When I started MailFinch, I did everything manually — printing, folding, stuffing, licking, stamping, addressing… everything. When I was ready to start scaling, I knew the pain points better than anyone else. Scaling was the easy part.