Rhapsody

We knew the big players like iTunes, Amazon, Yahoo Music and MP3.com were competing heavily on keywords related to “online music”, “buy music online” and “free music “. While the volume for these keywords was huge, we certainly didn’t have the ad dollars to compete on these search terms.

We used a tool called Keyword Spy, which tells us how much our competitors are spending on clicks (CPC). We then applied the average conversion rate of 2.5% provided by RealNetworks to determine the cost to acquire a new customer. Most competitors were running their paid search and display campaigns at a loss, paying upwards of $100 to acquire a free trial user. We had a $35 CPA to work with. We managed to get our CPA down to $15.

We did however notice a lack of competition on long-tail keywords—especially keywords related to artist, song and album titles (“refugee tom petty damn the torpedoes”).

There’s no way to test keywords without buying click, but we did have some tools at our disposal that helped us narrow down our original keyword set. First, we queried the Billboard charts to identify top selling artists and albums. From there we generated a list of song titles.

Next, we used the Rhapsody API to see which genres were being requested the most by current users. We found Jazz, Country, Folk, Classic Rock, Christian and R&B were the most popular genres for logged in users of the Rhapsody service. Once we identified these genres we used open source databases to retrieve artists, albums and keywords from each of these genres.

We now had research to inform two distinct keyword campaigns. One for top selling albums and their associated artists and songs and another long tail campaign containing keywords and ad groups for popular artist, albums and songs as requested by current Rhapsody subscribers.

In paid search, quality score can make or break your profitability. We knew we needed to have content on our landing pages specific to an artist and album. We also knew we needed to generate these pages quickly so we could test and exclude non-performing keywords ad groups.

We knew the Rhapsody API provided content and images for all artists in the Rhapsody database, so we built a templating engine capable of rendering 4-page artist specific microsite in a matter of seconds.

Each microsite contained artist bios, album and track information, images and a product tour. Oh, and a healthy dose of early-2000s web design flair. These microsites were all linked together, which increased our organic (SEO)...subsequently reducing our dependency on paying for clicks... which in turn lowered our customer acquisition costs. Cue fist bumps and happy hour.