Growth.

The framework

1. 01

   Put one monthly number on the people who don't pay you.

   You don't have to allocate this perfectly; you just have to stop leaving it blank. Add the infrastructure the free base consumes, the support tickets it opens, and the engineering and design hours spent each month on a surface only non-payers touch. Divide by the number of free accounts and you have a cost per free user per month, which is the price of the channel. Most operators can quote their advertising spend to the dollar and have never once produced this figure, which is how a business ends up running its largest acquisition channel entirely off the books.

2. 02

   The conversion rate is bad news in both directions.

   Under about one per cent and something is broken: either the free version is rich enough that nobody needs more, or the premium offer isn't legible. But Kumar's warning about the other end is the one people miss. A tier converting at half its users usually isn't a triumph; it's a free version too thin to pull much traffic in the first place, which caps the whole business at whatever that trickle amounts to. Read the level once to place yourself, then read the trend, because the level tells you far less than the direction.

3. 03

   Tune the boundary, the way the New York Times did.

   Kumar's example is a paper that ran unrestricted for years, then in 2011 limited readers to twenty free articles a month, decided it was still giving away too much, and cut that to ten in 2012. Two moves, one direction, both driven by the same reading — plenty of traffic, not enough subscribers. Expect to make that adjustment repeatedly rather than to get the line right at launch, and expect it to be unpopular, because people revolt at being charged for what they've grown used to receiving for nothing.

4. 04

   Today's rate will not survive tomorrow's base.

   Three mornings on comes channel decay proper — competition and habituation, forces outside your business. This one is internal arithmetic. Conversion is a property of the base's composition, not the product: widen it and you take in people with less need, so the share upgrading slides as the bill climbs with headcount. Apply today's rate to a bigger base and you count that growth twice. Kumar names LogMeIn and SugarSync as firms that met the squeeze and retreated to time-limited trials or dropped free. Against that sits a free user's other value: his research puts one at fifteen to twenty-five per cent of a premium subscriber, with a significant part of that coming from referrals.