home about contact internet marketing book twitter business book archives subscribe

Does Freemium Really Beat Ad-Supported?

In a recent poll, tech blog Mashable’s readers voted overwhelmingly in favor of freemium products over ad-supported products. Of the 1,250 readers who voted, 461 chose freemium as their preferred way of enjoying goods without paying for them, while just 305 would want more features but also lots of ads. That looks like a big thumbs-up then for products that give the basics for nothing, charge for the extras but keep advertisers away. But look a little closer at the figures, and the picture starts to change. The second-most popular option after “freemium” wasn’t “ad-supported.” It was “tie: both work.” The difference between that shrug and freemium came down to less than 40 votes. And that was a poll of users, not producers. If even customers are confused about whether freemium or ads are best, what are creators supposed to do? Should they be giving away the store but packing it with ads? Or does it pay better to give a little and charge for the rest?

Part of the confusion lies in the complicated nature of freemium. Chris Anderson, author of Free, a comprehensive look at the free economy, has identified four different kinds of freemium models: time-limited products allow users to try for a set period, commonly thirty days, before forcing them to pay; feature-limited products give away the basics but charge for more advanced features; seat-limited products let a small number of people use the same license but charge for mass use; and customer-type products let certain kinds of customers, such as small businesses, use the product for free, but charges those who can more easily afford it.

Cannibalizing Your Customers

Each of those models has its strengths and weaknesses. Time-limited products give users little opportunity to get to know the product well enough to find it invaluable, says Anderson. Customer-type products require difficult enforcement. And other models risk cannibalizing the low-end of the market by giving away the store to people who might have been willing to pay for it. Choosing between those models then looks difficult, but in practice, the nature of the product will play a big part in determining the most appropriate model. Seat-limited and customer-type products are only useful for software that might be used by office-loads of people, such as Intuit’s QuickBooks and Microsoft’s Bizspark. A feature-limited model will be of little use to a very simple program, such as a YouTube downloader; and a time-limited model would be a poor choice for a program that might be used to solve a one-off problem. Most producers then are likely to find themselves wondering whether to limit the features or restrict the time.

That choice should come down to money. Michael Mullany, vice president of marketing at Engine Yard, a ruby-on-rails app company, has tried to produce a mathematical equation for freemium models which, according to CNet means that:

freemium will be a better choice when:

Conversion Rate % > (Cost to Serve a Free User + Cost to Acquire a Free User)/Cost to Acquire a Paid User

Conversion rates for freemium programs tend to be between 2 and 8 percent. (Despite claiming that TurboTax online has a 70 percent conversion rate, Chris Anderson reckons that companies should target around 5 percent.) So for free or freemium to pay, Mullany concludes, free users have to cost between one-twelfth and one-fiftieth of the cost of picking up a paying customer.

But that’s comparing free or freemium to paid models. What happens when you compare freemium to the revenues that might come in through advertising, one particular kind of free model?

Ad Revenues are Unpredictable

First, things start to get more complicated because the revenue streams are more complex. While a paid business model will have a fixed income for each license sold, ad revenues can be fluid and unpredictable. A gaming company, for example, might give away the first level of its game for nothing, and have a 5 percent conversion rate of free users to customers willing to pay $10 for the full version. As long as it costs less than half a dollar to acquire each of those free customers, then the company will be in profit. But if the company is planning to make money not by persuading one in twenty to pay but by charging advertisers on a cost-per-click or cost-per-mille basis — or both — then it’s dealing with revenue figures that are a lot less clear than a sales price. The value of a cost-per-click will vary depending on the conversion rate and the price the advertisers are willing to pay for a lead. Cost-per-mille prices, too, can depend on the subject of the content. And the platform matters as well. Greg Yardley of Pinch Media, a mobile analytics firm, famously crunched the numbers and concluded that advertising on iPhone apps just doesn’t pay; even charging 99 cents, he argued, makes better sense.

Those low CPMs were partly down to a quirk in the iPhone. The lack of multitasking meant that clicking an ad yanked the user out of the app, making banners less effective than they should have been. The launch of a multitasking function in OS4, coupled with iAds, Apple’s new advertising network, has changed that situation. Large advertisers are reporting that they’re pleased with the network’s performance. Apple has now picked up $60 million in advertising commitments to see out the year and publishers are already feeling the benefits. According to the LA Times technology blog, Dictionary.com has managed to raise the price of its ad space by 177 percent since its enabled iAds in its iPhone app.

For developers then, choosing between freemium and ads still isn’t easy. There’s no overall strategy that fits every kind of product but rather some difficult cost and conversion calculations to compare against clickthrough rates and expected CPMs. Picking up all of that data won’t be as easy as it sounds either. Ask yourself whether you should go for ads or freemium and the best answer is likely to be the same as that given by users: a shrug and an “it depends.”

Leave a Comment