Wednesday, November 09, 2011 12:48 PM
Usefulness as a Pretense for Data Collection
Three years ago, at the Business of Software conference, I presented a Pecha Kucha (lightning talk) on how to gather data on your product’s trial usage when you aren’t a website.
I suggested that you should incorporate useful features into your desktop apps that need server support, and to do it in a way that is transparent to your users. In other words, don’t just contact servers with information secretly – instead provide online services integrated with your product, where the data you collect is a natural byproduct of the feature. The pretense for this was to find out who your active evaluators were, so that you could work to convert them.
So, whenever I see products that are not server-based start to incorporate online services, I wonder if it’s some kind of pretense to get data, and wonder why. After doing the iOS5 update on my iPhone, and using iCloud and iMessage and seeing the Siri demos, I started to see the pattern of online data gathering that Apple had started – which also includes Ping and Genius.
Each of these features uses online services in a way that is totally reasonable and understood by the user to be a natural way to implement these features. But, the end-result is that Apple is gathering a tremendous amount of data about its users. A dataset that may rival purely online companies.
Unlike Google and Facebook, Apple won’t use this for a targeted ad network (perhaps a little via iAd, but not in a big way). Instead, they will be more like Amazon and Netflix and use it to target content from their stores, and then as a basis for becoming more of a content producer/publisher.
For example, if you use Genius, Apple knows every song you own. Does it also know your ratings, and number of times played -- I don’t know, but I would want them to if it results in better Genius playlists. If you play your content out of iCloud, then Apple definitely has your actual consumption habits.
They already use this to market songs to you in the iTunes store. To take it to the next step, Apple could become a label (finding, nurturing and producing music), they’d have instant market research and reach. They already do this for apps.
Siri’s harder because it’s going to have a lot of odd data in it, but some of it is social graph related (who you call, meet, message and how frequently), which could certainly be used to influence what content you might consume (if all of your friends own the latest Lady Gaga album, perhaps you would like it too).
Another common pattern is a vertical search. Vertical searches are great to determine buying intent, which they could either service or broker through their stores.
Here are some sample requests shown in the Siri ad to get an idea of how Apple thinks you’ll use it:
- “Read me the message” (social graph)
- “Reply, ‘Definitely …’” (social graph)
- “Play my running mix” (music consumption)
- “What’s the traffic like around here” (vertical search)
- “Text my wife …” (social graph)
- “Is it going to be chilly in San Francisco this weekend? How about Napa?” (vertical search, consumption of travel)
- “How many cups are in twelve ounces?” (vertical search)
The advantage of using data to target content is that it doesn’t seem like an ad – you are only targeted when you go into the stores or are looking for content, and it just feels like a better, custom experience. Apple’s complementary moves are to then make as many stores as they can (iBooks, Newsstand, Apple TV), and then to start making the content themselves.
For physical goods, Apple could make brick and mortar store companions (like the Apple Store app), which help you buy at a store you are in – again, it won’t seem like an ad, just a guide, and it will be effective since being in a store is a great buying signal.
An even easier play is to make a Newsstand-like app for catalogs. “Siri, I need jeans”, “Siri, we’re out of coffee”, “Siri, help me get ready for bacon day next week”.