Blog Post

Case Studies in Freemium: Pandora, Dropbox, Evernote, Automattic and MailChimp

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Don’t spend money on marketing, do offer flexibility and data exporting to eliminate buyers’ regret, make sure to capitalize on and value goodwill, and only charge for things that are hard to do. That’s what some startups say is the key to success in the freemium business. But the biggest reason the five presenters this morning at the Freemium Summit in San Francisco — Pandora, Dropbox, Evernote, Automattic (see disclosure at the bottom) and MailChimp — are doing well is because they have great products that people want. They’ve been able to get those products to a broad audience by using the freemium model — that is, offering a free service with the option to upgrade. It’s an increasingly important business model, but one that’s hard to navigate, so their anecdotes, open sharing of data, and lessons learned were really valuable.

Pandora’s reverse-freemium approach: Pandora first launched in August 2005 with something that sounds quite similar to freemium: users got 10 hours of free online radio at signup, after which they were asked to pay $36 per year. “In the first couple weeks we had 100,000 people come through and the vast majority listened to every last minute of their free ten hours,” said CTO Tom Conrad. “Then we asked them for their credit card and they would wander off into the wilderness.”

Pandora CTO Tom Conrad
That November, Pandora switched on an “ad-supported” option. It was ad-supported in name only, however, because they had no ad server, no ad staff — not even a place on their page to put ads. But growth quadrupled overnight, and within three days, Apple called and asked to buy out ad inventory through December. Conrad and his team of course said yes, arriving at the price of $10,000 for the month. “We literally hard-coded the ads on the page,” he said. “We didn’t want them to know, but every time they changed their creative we’d have to relaunch the entire site.”

Pandora now has 20 million uniques and took in $50 million in revenue last year. Subscription rates had dropped to well below 1 percent of users. But 1 percent of a large number is still a large number, so last year the company launched Pandora One, a new take on premium with higher quality streams, a desktop app and fewer usage limits. It now has 300,000 subscribers, accounts for 1.6 or 1.7 percent of monthly uniques, and is expected to bring in 15 percent of 2010 revenue.

Dropbox’s numbers game: Dropbox CEO Drew Houston says you should know one thing about freemium: “It is a numbers game, so bust out your Excel spreadsheet. It’s all about finding things in the margins — lots of little things rather than one key thing.” Houston went into detail about the backup service’s attempts to recruit users through search marketing. The company found that obvious keywords like “online storage” were bid up, and the long tail of search terms had low volume. Then, people coming from search who actually signed up might not even pick the paid version (which has more storage and features).

Dropbox CEO Drew Houston
Ultimately that meant “our cost per effective acquisition per paid user was thousands of dollars for a hundred-dollar product.” So for a time, Dropbox went to great lengths to hide the free option to users coming in through search, and as a result confused users and felt terrible. “So the big lesson there is if you adopt a freemium business model your marketing cost is the free users.” The fact was that Dropbox was offering a product that people didn’t know they needed until they tried, and “search is great for harvesting demand, not creating it,” Houston said.

Having dropped search marketing as a strategy, Dropbox has actually grown incredibly fast. At some point, the company realized that user referrals were its biggest source of growth, so now it encourages referrals with an incentive program. That increased signups by 60 percent, said Houston, and it now drives 30 percent of total signups. The company now devotes 30 percent of its engineering to acquiring active users.

Houston also told a second story about Dropbox’s realization that its unlimited undo history — available to free and paying users — was responsible for a huge and growing share of its costs. And further, few customers actually used the feature. “We said ‘holy sh*t!’ More than half of our hosting costs are going to deleted, not restored prior versions of files,” said Houston. The company was wary of rolling back a free feature, but was able to manage the transition without too much fuss by telling customers about it openly, and giving existing users the option to keep the feature if they liked.

Evernote’s key metric: Evernote, the personal note-taking service, faces a challenge to spreading virally in that it’s not at all a social product, said CEO Phil Libin. But what Evernote can focus on is deriving maximum value from the users it does have. The company, which launched in June 2008, has 2.7 million users, with 7,000 new users per day (mostly through word of mouth, despite the lack of social features, said Libin), and 50,000 paying users (who convert in order to use the service on multiple platforms and for other premium features).

Evernote CEO Phil Libin
The thing is, over time inactive users drop off, and active users start paying. Once Evernote finally figured that dynamic out and started talking about it, term sheets and partnerships requests started flowing in, Libin said. “Our key insight is users are growing really fast, but revenue is growing faster.” Currently, users are growing 10 percent per month and revenue is growing 18 percent. “It’s like our users are a fine stinky cheese or wine — it gets better with age,” said Libin. “More and more people who aren’t going to pay just leave, and more who stay pay.” So 0.5 percent of people who sign up in a given month go premium, but 2 percent of people who signed up a year ago are now paying Evernote.

Libin’s key metric is comparing revenue per active user with variable expenses. At this point, the company makes $0.25 per month per active user, and spends $0.09 on variable expenses like infrastructure, customers service and network operations. He said freemium can work for any business if you have 1) a great long-term retention rate, 2) a product that increases in value over time and 3) variable costs.

Automattic CEO Toni Schneider and MailChimp CEO Ben Chestnut added a few more key lessons in their talks. Schneider talked about the blogging-software company’s decision to offer a-la-carte freemium services instead of tiered levels, giving both his team and users more flexibility. His company now makes 40 percent of its revenue from premium services like domain mapping, with the remainder from ad sales and enterprise products. But he said the problem with this approach is customers may not know of services they could receive, because it’s harder to market them individually.

Chestnut talked about the fact that free products are ripe for abuse. His 10-year-old email marketing company started offering a free version seven months ago, and has seen 240 percent user growth, a 225 percent increase in email delivery volume from 200 to 450 million, and a 200 percent projected revenue increase. But the biggest bumps of all? A 354 percent increase in abuse-related issues like spamming, followed by a 245 percent increase in legal costs dealing people trying to game the system. Luckily, MailChimp was able to develop automated ways to discover and deal with some of these issues, but even as an anticipated side effect the increase in abuse from going freemium has been huge.

Photos courtesy of Flickr users Hillary H, gaborcselle and hzeller, respectively.

Disclosure: Automattic, maker of, is backed by True Ventures, a venture capital firm that is an investor in the parent company of this blog, Giga Omni Media. Om Malik, founder of Giga Omni Media, is also a venture partner at True.

19 Responses to “Case Studies in Freemium: Pandora, Dropbox, Evernote, Automattic and MailChimp”

  1. Different tiers on Freemium aint working well. We also have success stories of youtube whose business model is now entirely based on advertisements. Pandora will have to pay 50 % of their revenues to the record labels as a royalty. 1 % of the total users as a subscriber group is very minimal. The key thing is whether freemium model is sustainable ? Freemium model must be changed to better respond to the traffic with brand loyalties and profit maximization. It will be interesting to see whether people still wanting to use Credit Cards over internet or they are causally using it ?

  2. Liz,

    Is any of the above mentioned companies making a profit? I know mailchimp does but how about the others? Are they living on venture capital? The thing with freemium is that requires very strong venture backup, not every start up can afford to have few million free users having uploaded 2GB of files on their servers. I believe based on my experience that freemium can stand as a strategy only if
    – you have significant funding from exprerienced VC’s or angels and not my ” rich uncle ”
    – Have a very clear path to exit the industry, lets be honest i do not think that is waiting to make few millions in profit because it will never do, it will very soon be acquired. On majority of these start ups the direction from investors is clear ” Grow as fast and as much as you can ” but where is the profit?

    Back in 2006 we run also an online file storage service, and we were on first page for 2 days, needless to say that we were getting 10,000 members a day but why? Because we offered everything for free, does not take much brain power to do that. However, smart people and teams are these who can convince people to pay because their product worth paying for.


  3. The freemium model is also in play in the online language learning space. Livemocha is disrupting traditional players (like CD-ROMs and language schools) with an self-study model that includes connections with native speakers. We’ve been adjusting and calibrating many of the variables mentioned here.
    The core challenge is striking the right balance between free features and content which generate viral growth and keep customer acquisition costs low, vs. “pressure points” and limitations that drive paid conversions.
    I think the best way to take on this challenge is to create an organizational mindset that is optimized to iterate rapidly and analyze the results. In fact, I’d suggest that this adaptive discipline is the defining success factor for a freemium model. Without it, you slow down and get risk averse, get locked into “stable metrics”, and fail to fully understand the levers that can propel the free footprint and revenue growth concurrently.

  4. These are great notes. I have some additional data on freemium and some takeaways from my time at the conference

    An excerpt:
    Generally, user conversion rates to paid products vary from 1-4% across sectors. Across small and medium business software, social and mobile gaming and music, paid upsell conversion rates were fairly constant. Many companies found that cross device usage of a product was the best correlating factor to paid upsell.

    Free users are assets, if you can control the cost: Free users build network effects and user mindshare that are useful for later upselling and refer a friend incentive program .Evernote measures revenue per active member per month as a key performance metric (KPI). In month 1, a user will generate $0.02 per month. In month 24, an active user wil generate $0.75. Focusing on upsell and building value drives enormous revenue opportunities. As for referral marketing, Dropbox’s two sided incentives offer increases conversion rates by 60%!

  5. Andrew Nadeau


    To your comment about “If these are all products customers really want/need and value then why not charge for it?”…

    I think freemium works really well when you have a product that people don’t realize how much they need it until they are using it. I use Dropbox to manage my files across different computers, clients, and co-workers. If Dropbox wasn’t free from the start, I probably would have never heard of it.

    If you have a great product that people know they need, then yes, you should charge for it from the start, which speaks to this quote from the article – “Search is great for harvesting demand, not creating it.”

    Freemium’s place in the world of business is for products that people don’t know they need, until they start using it. Once they use a product for free and see the value they will be willing to pay and will tell their friends.

    Giles, if you want to read more about the business models of freemium and SaaS companies then I recommend you check out David Skok of Matrix Partners extensive post on the topic at I think this is definitely a business model that is here to stay, and for us consumers is a great thing.

    • Andrew
      Thanks for the follow-up questions.
      Is it fair to say that the need must exist, whether or not customers recognize it or not*? The need is defined in the broadest sense and it is the job customers are employing the product for. After all the product was designed for serving a need.

      If we agree to this precondition, then it must be possible for a marketer to tease this out. Then it is also possible to create a favorable position for the product in the minds of the customer and make a credible value proposition (like it is done for all other products).

      In other words for any product there exists other ways for demand generation, implying freemium is just another way. But data is lacking on which method is superior.

      Note that even though customers may love the product and cannot live without it they may not be willing to pay for it. What would people pay for email?
      The problem with giving products for free is setting a bad reference price ($0) which affects the price people are willing to pay later. (See here for reference price effect Based on my long research into reference price I contend that people will not automatically start paying for something they used to get for free.

      I also do not think there is data behind the claim that WoM by initial users will drive further adoption. It might as well, but will it drive profitable growth? is it better than other marketing means?

      If one has done the analysis and finds that freemium offers the best of available ways for profit maximization then by all means one should do it. The trouble is choosing one over other without analysis behind the choice.


      *Of course, the need may be created because of the product and marketing as Kenneth Galbraith argued in The Myth of Consumer Sovereignity

    • It seems that word of mouth works for a lot of these freemium products. But with Evernote for example, (and I can personally attest to doing this), people sign up and then don’t use the product. If they’re signing up for the product and downloading the app, they’re showing some legitimate initial interest. It seems that its immediate value isn’t as apparent to some users.

      Check this post out – it discusses a great way to help eliminate this problem with an optimized conversion funnel mixed with behavioral ad targeting.

  6. Liz
    These are just three examples and write in absolutes without accounting for alternatives and the possibility that these cases could be exceptions and not the rule. It is easy to look at survivors, look at key observable characteristics and attribute their success to these. That is survivor-ship bias. What about businesses that followed the same principles of these success stories and did not make it?

    We should also note, despite what the article implies, that the freemium model is not the best of all available options. It is not clear to me whether the start-ups have evaluated all available options and decided freemium was the profit maximizing route.

    The conversion from free to paid versions are neither predictable nor reproducible. The businesses did not have set themselves a target rate before they launched the freemium model. I do not understand how quoting this metric will serve others considering this model. Getting customers to sign up for the free version and hope a percentage of them will upgrade to paid version is not strategy – customers can remain happy with the free version longer than businesses can stay solvent (hat tip to Keynes).

    If these are all products customers really want/need and value then why not charge for it?

    • Free Fall

      The essay is not actually promoting the freemium model. It is merely providing a four pronged case study of where it has worked.

      Also, while the conversion from free to paid is not predictable at first, neither is whether ads can create a substantial demand. Once the process has started, statistical analysis shows trends for both.

      Why is it not a valid strategy? It is a risky one, I’ll give you that, but it is still a strategy.

      To get any real benefit from following these businesses, one would have to go far deeper than the scope of the article.

      In short: your arguments are outside the scope of the article.

    • abhiroopb

      I disagree “trialware” generally only allows you to use a product for a limited period of time. In some cases an essential feature is blocked of (for example Partition Magic’s trial version doesn’t actually allow you to partition!).

      The freemium model allows you to use a product completely for free only paying a “premium” for certain extra features (generally for the power user).

      • But when Pandora offered a trial they didn’t get uptake. One thing that freemium does (and yes, perhaps it’s just how you word it) is give customers the idea they are getting the full service for as long as they want without pressure to pay unless their needs change. Then, they can upgrade or downgrade for added features as they see fit.

    • You are not wrong, but this is the evolution of the term and in many ways, the industry too. In his book Free, Chris Anderson defines several methods of freemium such as Time limited, Feature limited and Capacity limited. All of which amount to trialware as you’ve said. The difference is today the free is an incentive to buy the premium, where once it was a trial of the product itself. The marketing tactics, and indeed the industry, has evolved :-)

  7. It would be interesting to see a discussion, not only of the value of each of the popular freemium services in isolation, but as a group. For example, using EverNote to make quick notes on a mobile device, and saving it to a dropbox folder for future use. There are probably many other combinations as well. Makes you wonder if people would pay for the value of the bundle if they had to make a choice.

  8. Email Marketing does not have to mean mass email!
    Companies invest a great deal in their website which in many cases is their only “store” where they showcase products and services. The challenge is to drive people to the website but as we all know, more people in the store will lead to more sales. Corporate employees send emails every day to clients, prospects, friends and others but these are plain emails that do not generate any traffic to the website.
    WrapMail – – offers a solution that does not require any installation but that seamlessly adds interactive letterheads (designed by the client) to every outgoing email so that each and every one becomes a promotional piece for the company and when clicked delivers the reader to the website. Furthermore the solution tracks the clicks and reports who is clicking on what and when (also in real time), turning the system into a research tool.
    This “hidden” advertising medium is probably the most viral available and the least costly, WrapMail only charges $5 per user per month.