Google's latest announcement of two new social apps - 'Allo' for direct messaging, and 'Duo' for one-to-one video - immediately brought back memories of the company's biggest foray into social media in Google+. For those who don't understand why Google+ failed to become a significant player in the social space, here's a creative depiction of how it went down.
Google: "Hey guys, we've got this really cool new social app - better than Facebook. It's got all these new features and advances that are well beyond..."
Users: "It does all the same things we can do on Facebook."
Google: "Yes, but better. We've..."
Users: "Whatever nerds. Poke us when you've got something original."
That's obviously an over-simplification, but you get the gist. Google+ tried to beat Facebook by being better. But how Google defines 'better' and how users do is actually very different.
Google+ does have a range of features that are technically impressive, but essentially, there wasn't enough differentiation in Google's social offering to inspire a large enough number of users to leave the platforms on which they'd already established a presence and migrate across to make Google+ their new social home. Even when Google tried to force people to try it out by making it mandatory to have a G+ account to comment on YouTube videos, even when they made it an essential app for brands by giving those on G+ an SEO boost, people just didn't take to it. And eventually, it just fizzled out.
But Google knows that social is important. Google sees the ever-growing threat of Facebook looming, they know they need to adapt to changes in user behavior in order to maintain relevance.
For example, Google searches via mobile surpassed desktop for the first time in 2015, which underlines the overall trend of users moving towards mobile devices. And on mobile, both the Facebook and Facebook Messenger apps ranked above Google search in unique users last year.
Google's also been very secretive about how many mobile searches are being conducted via their platform, which may, possibly, suggest that that number is in decline. In a post on Search Engine Land last year, Danny Sullivan noted that the most commonly shared figure on how many Google searches are conducted per day, in total, is 3.3 billion. But that figure comes from stats released by Google in 2012 - since then, there's been no official update on total Google search volume. Independent research, however, has suggested that mobile search volume could be a significant concern for the search engine, with the average mobile user conducting less than one Google search per day.
In addition, Facebook says it's already facilitating more than 1.5 billion searches on its platform every day, an area Mark Zuckerberg sees as "a huge opportunity to create value". Facebook's also working to become the place where people can do everything, without needing Google, by building a new bot platform to enable people to find more info within Facebook Messenger - which, itself, has now passed 900 million monthly active users.
Google can see the trends, they know that social's an important part of staying connected and delivering on audience needs, and that messaging, in particular, looks like it could become a major headache if they don't get on board. So they have to do something.
I imagine the pitch for Google+ rested on a similar impetus.
Similar, But Better
But just as many of the technological advances of Google+ were overlooked because it lacked differentiation, the same could be set to happen again with Allo and Duo. This is most notable with Allo - in amongst the grumblings about Google releasing yet another messaging app into an already crowded marketplace, many people seem to have already overlooked the advanced image recognition AI capabilities on offer within the app.
The system works like this: When someone sends you a photo within Allo, the system will scan it, using Google's advanced image recognition AI. From there, Allo will not only be able to determine what's in the image, but it'll also be able to formulate your most likely response, based on your interactions with similar content and what Google's Knowledge Graph understands about the image subject.
Like Facebook, Google's been working on their image recognition AI for some time, training their systems to develop an understanding of image inputs in order to provide a more accurate, automated description of image content.
Using this, Allo aims to make it easier for people to respond to image content on the go, when you don't have time to be thumbing in responses. It does this by developing an understanding of the various contextual markers at play - for example, Google's AI knows that spaghetti and linguine are visually similar, so it takes into account your responses to both when making suggestions for how you might want to reply to an image.
Through this, Google's system's able to suggest responses based on the image content. So rather than just your basic descriptors around what's in a photo, which can seem quite wooden and manufactured.
The system can also suggest likely responses based on the actual context.
The image recognition AI and response will develop over time as more people use it, but it is something of a technical achievement already, even at this stage of the program. Of course, Facebook's already rolled out their image recognition AI to automatically generate image descriptions for visually impaired users, and it's hard to know which company is more advanced on this front. But through Allo, Google's making this function a key feature, a crucial differentiator within the app.
Which is great, but the question is whether users will really care.
If they system works as advertised and it provides intelligent, intuitive response for users to chose and response to images, it could be an interesting novelty that might draw more users to the app. But then again, maybe it's too artificial, maybe people won't warm to automated responses, even if it is more efficient. It does seem like a step back from the connectivity of social, as opposed to a new way to stay closer to your connections.
But either way, the advances in image recognition that we're seeing are amazing, and they have the potential to change the way we monitor and respond to social media cues.
Visual Response
The best thing about the evolution of image recognition technology is the opportunities it'll bring.
This is why it's important we don't overlook this element of Allo, or indeed of Facebook's image recognition efforts - at some stage, image recognition will open up new doors and provide new data sources to work with which will change the way we track and respond to social media mentions.
For example, you'll soon be able to set up social media monitoring not only for keywords and mentions, but also for visual cues - branding in photos, products. That might not seem like a major element to factor in, but when you consider that there are more than 8,700 photos shared on Snapchat every second and more than 400 million photos shared across Facebook and Instagram every 24 hours, that's a huge data pool that's currently not able to be tapped. And that's before we even consider videos, which are constantly growing.
Being able to monitor for visual cues adds a whole new element, and one that could prove extremely valuable in years to come. As noted, it'll take some time to develop, but seeing the capacity of Allo, it may not be that far off.
Through visual monitoring, you'll be able to:
- Target users with ads when they're near your store - even if they're not geo-tagging their posts, you could use visual recognition to match their location
- Target promotions to people who already buy your products, as the image data would show where your products have appeared in their photos, and, ideally, identify if they were in the background or foreground to add extra contextual understanding
- Reach out to users who not only mention your content, but also show it in their photos and videos. Right now you can get in touch with those who mention your brand, but visual monitoring would enable you to locate real brand advocates by cross-matching mentions and visual references
- Get a better understanding of consumer based not only on what they say, but also what they do. Someone who regularly posts images of their car would be a good target for car care products, someone who regularly posts images of their baby would likely be interested in toddler fashion trends. We can determine all this, at least to some degree, with the data we have now, but image data will provide a heap more context, further solidifying your outreach planning
There's a whole range of other ways image recognition data will help. And while it'll also come with a new set of privacy concerns, the development is huge, and really worth paying attention to.
It's impossible for us to know, at this stage, whether Allo and Duo will take off, but it's important to recognize the technical capacity at play, the advances that are being rolled into these new products and what they mean for the future of social media marketing. Just as Google+ had some great, valuable elements that were widely over-looked, Allo may too fail to catch on, despite its capacity. But don't overlook where that technology is at, and where it's headed.
And while Google may not have found it yet, you can expect they'll keep searching a solution to stay in touch with mobile and social trends. Because really, the data shows they have to. Google's revenue is obviously very strong, it's not an immediate threat to their business. But over time, Google knows they need to adapt. One way or another.