for baber makayla only

Pandora is the Internet’s most successful subscription radio service. In January 2011, it had over 80 million registered users in the United States, and continues to add about 600,000 new subscribers a week—that’s one new subscriber about every second! Pandora now accounts for over 50% of all Internet radio listening hours. Radio? In the Internet age of iTunes, Rhapsody, and listen-to-what-you-want-anywhere-anytime? Why would anyone want an online radio station to choose the music they will be able to hear? That’s so old school.

Not exactly. At Pandora, users select a genre of music based on a favorite musician, and a computer algorithm puts together a personal radio station that plays not only the music of the selected artist but also closely related music by different artists. How does the computer know about closely related music and music genres? Can a computer understand music? Not really. Instead a team of professional musicians listens to new songs each day and classifies the music according to over 400 musical criteria including male or female vocal, electric vs. acoustical guitar, distortion of instruments, presence of background vocals, strings, and various other instruments. These criteria are used in a computer algorithm to classify new songs into five genres: Pop/Rock, Hip-Hop/Electronica, Jazz, World Music, and Classical. Within each of these genres are hundreds of sub-genres. Like Taylor Swift? Create a radio station on Pandora with Taylor Swift as the artist and you can listen all day not only to some Taylor Swift tracks but also to musically related artists such as Carrie Underwood, Rascal Flatts, Anita Nalick, and others.

The algorithm used to identify genres of songs is a result of the Music Genome Project conceived by Will Glaser and Tim Westergren in 1999. Westergren, a jazz musician, and Glaser believed it was possible to identify genres of music, and sub-genres, using their expertise (and that of other musicians) to identify similarities among artists and songs. They have identified over 400 factors to help classify songs, and leave it up to the computer program to select appropriate matches based on a user’s input of a selected artist. To some extent they are mimicking disc jockeys and radio program managers who had no trouble creating jazz radio, classical radio, and pop/electronica stations, and within these general categories, sub-groups of musicians who shared musical characteristics.

Top of Form

In 2005, Glaser and Westergren launched

Pandora.com

, a music service based on the Music Genome Project. Their biggest challenge was how to make a business out of a totally new kind of online radio station when competing online stations were making music available for free, most without advertising, and online subscription services were streaming music for a monthly fee and finding some advertising support as well. Actually, their biggest challenge was to avoid going broke: over 80% of online music is downloaded from P2P networks for free. iTunes launched in 2001 and by 2005 was a roaring success, charging 99 cents a song with no ad support, and 20 million users at that time. The idea of a “personal” radio station playing your kind of music was very new.

Facing stiff odds, Pandora’s first business model was to give away 10 hours of free access to Pandora, and then ask subscribers to pay $36 a month for a year after they used up their free 10 hours. Result: 100,000 people listened to their 10 hours for free and then refused to use their credit cards to pay for the annual service. People loved Pandora but were unwilling to pay for it, or so it seemed in the early years.

Facing financial collapse, in November 2005 Pandora introduced an ad-supported option. Subscribers could listen to a maximum of 40 hours of music in a calendar month for free. After the 40 hours were used up, subscribers had three choices: (a) pay 99 cents for the rest of the month, (b) sign up for a premium service offering unlimited usage, or (c) do nothing. If they chose (c), the music would stop, but users could sign up again the next month. The ad-supported business model was a risky move because Pandora had no ad server or accounting system, but it attracted so many users that in a few weeks they had a sufficient number of advertisers (including Apple) to pay for their infrastructure. In 2006, Pandora added a “Buy” button to each song being played and struck deals with Amazon, iTunes, and other online retail sites. Pandora now gets an affiliate fee for directing listeners to Amazon where users can buy the music. In 2008, Pandora added an iPhone app to allow users to sign up from their smartphones and listen all day if they wanted. This added 35,000 new users a day. By 2009, this “free” ad-supported model had attracted 20 million users. All of Pandora’s plans come with restrictions required by the music companies that own the music, including the inability to hear a song on demand, no replay, and a skip limit of six skips per hour per station. Also, the music cannot be used commercially or outside the United States. After struggling for years showing nothing but losses, threatened by the music labels who wanted to raise their Internet radio rates, Pandora finally had some breathing room.

Still not giving up on its premium service, in late 2009, the company launched Pandora One, a premium service that offered no advertising, higher quality streaming music, a desktop app, and fewer usage limits. The service cost $36 a year. By July 2010, Pandora had 600,000 subscribers to its premium service, about 1% of its then 60 million users. At the end of 2009, Pandora reported its first profitable quarter and $55 million in annual revenue, mostly from ads. The remainder of its revenue came from subscriptions and payments from iTunes and Amazon when people bought music. In 2010, Pandora achieved even greater success. Revenue more than doubled, to $137 million, with about $120 million coming from advertising and $18 million from subscriptions. Pandora’s “new” business model has proven so successful that it filed for an initial public offering in early 2011, and went public in June 2011 with a valuation of $2.6 billion!

Pandora is an example of the “freemium” business revenue model. The model is based on giving away some services for free to 99% of the customers, and relying on the other 1% of the customers to pay for premium versions of the same service. Chris Anderson, author of Free: The Future of a Radical Price, noted in a blog post “…you give away 99% of your product to sell 1%. The reason this makes sense is that for digital products, where the marginal cost is close to zero, the 99% costs you little and allows you to reach a huge market. So the 1% you convert, is 1% of a big number.” There are many other examples of successful freemium model companies. For many traditional print media like newspapers and magazines, the freemium model may be their path to survival. But it won’t work for every online business.

While it clearly has worked for Pandora, there is ongoing debate among e-commerce CEOs and venture capitalists about the effectiveness of the freemium model. The crux of the issue is that while freemium can be an efficient way to gather a large group of potential customers, companies have found that it’s a challenge to convert eyeballs into those willing to pay. Absent subscriber revenue, firms need to rely on advertising revenues.

MailChimp’s story is both a success and a cautionary tale. The company lets anyone send e-mail newsletters to customers, manage subscriber lists, and track the performance of an e-mail marketing campaign. Despite the powerful tools it gives marketers, and its open applications programming interface, after 10 years in business, the company had only 85,000 paid subscribers.

In 2009, CEO Ben Chestnut decided that it was time to implement new strategies to attract additional customers. MailChimp began giving away its basic tools and charging subscription fees for special features. The concept was that as those customers’ e-mail lists grew, they would continue using MailChimp and be willing to pay for enhanced services. These services included more than just the ability to send e-mails to a greater number of people. Clients would pay to use sophisticated analytics to help them target their e-marketing campaigns more efficiently and effectively.

In just over a year, MailChimp went from 85,000 to 450,000 users. E-mail volume went from 200 million a month to around 700 million. Most importantly, the number of paying customers increased over 150%, while profit increased over 650%! Sounds great, but there was also a price to pay. The company also saw a significant increase in abuse of its system, and a related increase in legal costs. Much of the abuse was “fuzzy spam,” where spammers were able to successfully disguise their efforts so software such as Spam Assassin couldn’t find and block their messages.

Instead of abandoning the freemium model, MailChimp developed Project Omnivore, an optimization algorithm that could find bad e-mails. Between September 2009 and September 2010, the company sent around 70,000 warnings, suspended almost 9,000 accounts, and shut down about 1,900 users. Fortunately for MailChimp, the algorithm can add additional value for customers because it can find positive trends and estimate the odds that users will open a given e-mail in an e-marketing campaign.

For MailChimp, freemium has been worth the price. It currently supports more than 900,000 subscribers worldwide, sending 1.75 billion e-mails per month. However, Ning, a company that enables users to create their own social networks, tried freemium and came to a different conclusion. They abandoned it in July 2010.

Marc Andreessen, co-author of Mosaic, the first Web browser, and founder of Netscape, launched Ning in 200

4.

With his assistance, the company has raised $119 million in funding. Despite being the market’s leading social network infrastructure platform, Ning was having a common problem—converting eyeballs into paying customers. While 13% of customers were paying for some premium services, the revenue was not enough. The more free users Ning acquired, the more it cost the company.

In May 2010, Ning announced the impending end of the freemium model. The company shed staff, going from 167 to 98, and is now using 100% of its resources to capture premium users. It has introduced a three-tiered pricing model, starting at $19.95 per month for the basic service. Ning gave existing users a 30-day grace period to choose a plan. So far, about 50,000 Ning networks have converted to paying status, over three times the number of premium sites (15,000) it previously had. Ning also said it was adding new paying subscribers at the rate of 5,000 a month, which is also triple the previous rate. Revenue is expected to be up by around 400%, and the company expects to be profitable by the end of 2011.

Ning’s new business model counts on a variety of revenue streams: premium services, Google ads, partnerships, and sponsorships. Partners include Pearson Publishing, which has committed to sponsor Ning’s 8,600 education networks (K-12 and higher education) for the next three years, reducing the cost to subscribers to just $19.95 a year. Additional partners include CafePress, which helps customers sell branded merchandise online, and ChipIn, which helps nonprofit customers raise money and collect donations.

Top of Form

So when does it make sense to include freemium in a business plan? It makes sense when the product is easy to use and has a very large potential audience, preferably in the millions. A solid customer value proposition is critical. It’s helpful if a large user network increases the perceived value of the product (i.e., a dating service). Freemium may work when a company has good long-term customer retention rates and the product produces more value over time. An extremely important part of the equation is that the variable costs of providing the product or service to additional customers for free must be low.

For example, Evernote, a personal note-taking service, added freemium to its business model and grew its user base to 8 million. Around 25,000 new users sign up every day. The company has over 300,000 paying users. Typically, 2% to 5% of freemium users convert from the free product to the paid version. Ever-note currently has a conversion rate of around 3%, within the range of what is expected. But Evernote has also discovered that the longer a subscriber remains an active user, the more likely he or she is to convert to a premium subscription. For instance, about 20% of the subscribers who have used Evernote for two years have become paying customers. The company is now generating revenue of about $1 million a month. One of the key metrics for success for Evernote, and any other company using freemium, is revenue per active user versus variable expenses. Evernote earns $.30 per user per month, and it spends only $.09 per user per month.

Companies also face challenges in terms of what products and/or services to offer for free versus what to charge for (this may change over time), the cost of supporting free customers, and how to price premium services. Further, it is difficult to predict attrition rates, which are highly variable at companies using freemium. So, while freemium can be a great way to get early users and to provide a company with a built-in pool for upgrades, it’s tough to determine how many users will be willing to pay and willing to stay.

A freemium strategy makes sense for companies such as Pandora, where there is a very low marginal cost, approaching zero, to support free users. It also makes sense for a company where the value to its potential customers depends on a large network, like Facebook. Freemium also works when a business can be supported by the percentage of customers who are willing to pay, like Evernote and Pandora, especially when there are other revenues like affiliate and advertising fees that can make up for shortfalls in subscriber revenues.

1.

Compare Pandora’s original business model with its current business model. What’s the difference between “free” and “freemium” revenue models?

2.

What is the customer value proposition that Pandora offers?

3.

Why did MailChimp ultimately succeed with a freemium model but Ning did not?

4.

What’s the most important consideration when considering a freemium revenue model?

Bottom of Form

Bottom of Form