Are App Graphs The Next Big Thing In Search?
You've heard of social graphs and knowledge graphs, but what about an app graph? Columnist Daniel Cristo explains how this may (soon) be a factor in search.
This week, Twitter announced that it will be tracking which mobile apps users download in order to improve its ad targeting capabilities.
If you use Twitter’s mobile app, you’ll soon be presented with a prompt inside the app that reads, “To help tailor your experience, Twitter uses the apps on your device.”
You can choose to go into your settings and opt out, or you can ignore it and allow Twitter to see which apps you download to your device. With this information, it will create an App Graph, which helps the company better understand user behavior and interests.
What’s An App Graph?
Maybe you’ve heard of Facebook’s Social Graph and Google’s Knowledge Graph, but what exactly is an App Graph?
Graphs are a way of structuring information. For example, if I showed you a random person’s phone, and on that phone was a fashion app, a budgeting app and kids’ games, you’d probably assume the user was a young mom.
On the flip side, if the phone had a couple of business and finance apps and a bunch of first-person shooter games, you might assume the user was a young male right out of college.
While you do have to make some inferences regarding users, apps can be quite a rich source of data for companies looking to enhance customer profiles.
The App Graph & Search
If Twitter thinks mining your app downloads improves its advertising capabilities, and Facebook is already doing this with apps that use its developer’s kit, you can be sure that Google hasn’t ignored this opportunity — especially considering that it owns the Android platform and it has some of the most popular apps on the iOS platform.
I don’t think the question is whether or not Google has its own App Graph — the company would be foolish not to. The real question is whether it is using that information to improve organic search. I tend to think it probably is, and here’s why…
1. Google Already Personalizes Results
If personalizing results leads to a better search experience, and an app graph helps Google better personalize the results, then it just makes sense to use that data.
Accessing the data should be easy; but matching it back to a Google account and working it into its algorithm is a little harder.
2. Apps Have A Great Signal-To-Noise Ratio
Let’s say you want to use Facebook “Likes” as a means to understand what someone actually likes. When you look at the data, you’ll see that people do actually “Like” some of the people and businesses they like in real life, but they also “Like” a lot of junk that they would never endorse offline.
That junk is called noise. It makes it hard to put faith in Facebook “Likes” as a credible reflection of real life.
Mobile apps, on the other hand, are different. You only have so much room on your phone for apps, so you need to be more selective of what you download. If you have an app that isn’t being used, eventually you’ll delete it to make room for more useful ones.
Downloaded apps are also more private than social actions like Facebook “Likes,” and that perceived privacy allows for a little more honesty.
Sure, there are probably a few apps we download out of curiosity that really don’t match our interests or personality. That said, it’s unlikely we’ll use those apps frequently. So filtering by frequency (and filtering out the noise) should give an app graph a lot of accuracy.
3. It’s A Competitive Advantage
Google’s only real search competitor, at least in the U.S., is Bing. Bing is owned by Microsoft, which has its own mobile platform.
Unfortunately for Microsoft, its Windows Mobile devices have a much smaller share of the mobile market than Google. Google’s market penetration gives Google’s Android platform a much larger and much richer app graph than Microsoft’s, which should result in a competitive search advantage if the app graph was used as a search signal.
4. It’s Very Hard For Spammers To Manipulate
One of the most challenging problems Google faces is spam. Once webmasters figured out how important links are to search rankings, they spent a lot of time and energy figuring out how to create artificial links to their sites, which manipulated search rankings.
Apps are a little bit different. Their home screen is a very personal space, and bribing someone to download and use an app is much harder than a one-and-done link or share. Additionally, it would be incredibly expensive to try and do it at scale. You’re just better off creating an app people actually enjoy using.
5. Nobody Is Looking To Apps As A Signal — Yet
We all saw what happened to the SEO space once it became common knowledge that links were a strong search signal — everyone started buying links. This is exactly why Google doesn’t publish its algorithm. It needs to keep some parts of it a secret, otherwise spammers will have an easier time gaming the system.
Therefore, the best ranking signals are the ones nobody knows about. That’s why I call BS on Google announcing signals such as HTTPS. Once you tell people, “Hey, this is going to affect your rankings,” they stop doing it naturally, and you’re left with a lot of useless noise.
If Google really intends to use the App Graph as a search signal, it will keep quiet about it as long as it can.
It seems to me that app graphs can provide Google with a new layer of personalization, which when combined with its existing social graphs and interest profiles, could improve the relevance of both paid and organic search.
What To Do?
The best advice I can give you is to help Google understand who your site is for, as opposed to what keywords your site should rank for. Instead of trying to rank for “luxury cars,” focus less on writing content that talks about luxury cars and more on content that appeals to the lifestyle of the person who buys a luxury car.
In this way, Google begins to understand that your site is geared toward a certain type of user. When that user searches for related content, Google should give your site a ranking bonus. App Graph data plays its part in helping Google understand the user better.
The App Graph is just another instance of how search engines and social networks are blurring the lines between their core businesses. Search engines are using app graphs and social graphs to move search away from optimizing for keywords and toward optimizing for user profiles, while social networks are using big data algorithms to connect users with relevant people and content.
This shift should be a good thing for marketers which are able to get beyond the single-channel paradigm to a more holistic approach to inbound marketing.
Share your thoughts on app graphs below, and don’t forget to opt out of Twitter’s app tracking; unless of course, you like that kind of thing.