Hitmetrix - User behavior analytics & recording

Music and Massages: Voices from Dreamforce #3

What comes to mind when you think about Pandora? Streaming music, of course. Perhaps personalized Internet radio, where the next track is always being selected for you, based on your prior listening habits? A subscription service?

All correct up to a point. But none of that goes to the heart of the matter. In a conversation at Dreamforce with Dave Smith, Pandora SVP for monetization and yield, I repeatedly heard the brand described as a publisher. Indeed, Pandora’s business model is about as much about selling you music, as Vogue’s is about selling you glossy magazines. “About 90% of our revenue comes from advertising,” he told me. Pandora is a publisher, in the sense that its business is to “accumulate audience and make it available to advertisers.”

The reason it’s been able to do that so successfully is — you guessed it — data.

Pandora’s audience is an audience of registered members, of course, which means the brand is working with deterministic data: email address, birth year, zip code, gender. “We audit the heck out of age and gender,” said Smith. And then there’s the overlay of inferred data from listening habits — ethnicity (habitual listeners to Hispanic artists, for example), children in the house, location (85% of listening takes place on mobile). “We don’t know what kind of toothpaste you use or car you drive, but there are companies which sell that data, and we’re a pretty material buyer.”

Pandora is at Dreamforce because it’s a user of the Marketing and Sales Couds. As for sales (“I don’t own that part of the relationship,” Smith said), Pandora has been a Salesforce CRM user for “a long time.” Pandora, of course, is B2B-facing too, selling streaming to “large broadcast buyers” like Home Depot. The Marketing Cloud component which is front-of-mind for Smith is the DMP formerly known as Krux, “the pipes” for the third party data Pandora purchases. It pushes its own first-party and inferred data into Salesforce DMP; the DMP appends anything else it might know about that customer. “Salesforce manages the connectivity of that data out into the eco-system.”

Performance-oriented advertising

Pandora’s sweet spot is performance-oriented advertising: ads which seek to drive specific behaviors like app downloads or “hard core” conversion. Smith cites AT&T as an advertiser seeking true direct response outcomes. The market includes both large national advertisers, and local advertisers looking for foot traffic. There, Pandora is competing with local radio stations. The data feedback loop is closed by looking at audience members’ behavior before, during, and after exposure to a campaign.

About 40% of all the ads Pandora sells are touched by this data, and there’s “an explicit price uplift” — these are premium segments. Advertisers pay “more than I pay for the data,” Smith chuckles. For the top tier of strategic advertisers only, Pandora is starting to license its data for use outside the Pandora platform. This is just for key partners where relationships are well-established and growing.

Data at the core

“Data has always been at the core of our content offering,” said Smith. Pandora is known, of course, for its uncanny — but entirely explicable — ability to predict “next best song” for any listener. The brand’s music genome project is an ambitious attempt to categorize the world’s music by hundreds of attributes. While some attributes — BPM or key, for example — can be tagged automatically, the project is essentially one being undertaken by trained human listeners. They’ve analysed some 25 million tracks so far, Smith told me, using the human ear to understand things like tone and mood.

Most of the competition — Smith says — are working with well-filtered genre lists. “We’re much deeper than that.”

That prompted me to ask whether it was true that a streaming music service like Pandora can track listeners’ sentiments around the clock through playlist trends. “We haven’t figured out how to infer mood,” Smith admitted. Also: “We’re very careful about impinging on privacy.”

Massage Envy: National Profile, Local Voice

Scottsdale, Az.-based Massage Envy offers not only customized massages, but also facials and a range of skin care treatments, on a franchise model. Indeed: “Two-thirds of our marketing spend is managed locally,” Jeanna Corley, the brand’s division VP of integrated marketing, told me.

That makes Massage Envy something of a natural for the kind of distributed marketing model Salesforce is now offering through the Marketing Cloud. 

Although its a franchise-based operation, Massage Envy’s objective, Corley said, is to “be the national brand we should be.” And they’re planning to get there by improving efficiency in their marketing, and by leveraging data. Raising the brand’s profile is not just about customer acquisition and retention. “It’s key to our growth to attract more massage therapists,” she said.

Building out a platform to achieve this meant interviewing a number of potential partners, Corley said. “We brought in several and talked to them. Salesforce was a natural fit” — not least, she said, because she feels Massage Envy can grow into Salesforce’s offerings.

They started simple with email: building the segments, building the customer journey, applying attributes. It took 90 days to get to the first non-fancy email piece. Corley considers this fast: “It had to be, because we’d sold the idea to the franchises.” Corley worked at first with hand-picked franchises from just under 700 locations, “who would work with me as my partner.” The franchise connection is important: Although the aim is to leverage the brand “from a national perspective” while retaining “a local voice.” She plans custom templates for local campaigns.

The next channel of interest is SMS. “I’m also interested in some of the automation pieces, when we’re more into our run phase, when we get a bit further.”

Massage Envy did use an implementation partner, Levementum, not least to get their data in order. It was distributed across a large number of databases, and one of the main problems was “de-duping.” A customer who visited a number of franchises could have multiple records across the whole set. Internally, Corley has a CRM manager who is the Salesforce expert, and is looking at adding resources under him.



Salesforce covered DMN’s expenses to attend Dreamforce 2017

Total
0
Shares
Related Posts