The time for ‘Visual Discovery’ is nigh

What do I mean by ‘Discovery’?

The way in which people ‘discover’ information and content online has been essentially owned and shaped by Google. This is true for the majority of global internet users since the search engine rose to prominence in the late 90s and early 2000s.

The idea being that if you want to find something, you either visit (or more likely already have access to on whatever page you’re on) Google’s search engine and type in the words that explain what you’re looking for. Simple right?

Very early in its history, Google understood the commercial implications of the results it served up, particularly in relation to any search terms that may reference specific products or suggest that we, as users, want to purchase a specific item. It’s this type of commercial search that I’m going to focus on in this article, asking whether there are now real world examples of when we can do better than words.

Take the search below:

In the search above, everything at the top of the screen and the first result served up (note that it’s even marked as such) is paid for by advertisers. It’s this kind of advertising revenue that drives the majority of Google’s revenue (in truth there’s a whole range of different advertising revenues that Google taps, but for the sake of the simplicity of this article I’ll simplify!)

Now — the above customer journey works just fine assuming certain things.

  1. I already know the exact product / service I’m looking for
  2. I know the specific name of the brand / provider of that product / service.
But what if I don’t?

Without those two factors, typing into Google, as fantastic, as brilliant and as incredible a tool as it is (it really, really is!)… this experience actually feels imperfect. It feels dated compared with the range of technologies now better suited for this kind of inspirational search or online ‘window shopping’.

Imagine I was looking for the same shoes from the example above. Imagine I’d seen people wearing them, I really liked them too and maybe I can even picture them in my head. But for the life of me, I just can’t remember what brand they are or what they’re called.

I bring up Google and… I’m at a loss.

I can only describe them imperfectly and Google’s algorithm can only guess what I’m most likely to be looking for based on factors like how other people have behaved after searching for that term… But ultimately, I don’t find what I was looking for. :-(

Enter Visual Discovery

Many of us over the years will have just grown accustomed to the above being how the internet works, without ever really asking what the potential alternative is. But words feel clumsy and imprecise in this kind of example, and we now have the technology to do better.

There’s an old saying that ‘a picture is worth a thousand words’ and it feels apt to wheel out that old chestnut now (it’s cool to mix idioms right?). Well, it does really feel like there are plenty of scenarios where images, pictures and communicating visually is far more effective than having to use written (or even spoken!) words.


Compare the above customer journey with the direction that Pinterest’s newly released ‘Visual Search’ tools are heading.

Through ‘Pinterest Lens’, ‘Instant Ideas’ and ‘Shop the Look’ (read more here) users are provided with the ability to receive related results based on photos from their phone’s camera (Pinterest Lens), any image you find online (using the Instant Ideas Chrome extension) or buy the items directly from images (Shop the Look)…

In practice, Pinterest Lens is still a little way from being at the point of recognising specific models and brands, but as the algorithm is refined I have no doubts it will get to that point. It’s already a fantastic tool for inspiration or finding comparable products to ones in front of you — example below.


Blippar repositioned itself in mid-2016 to also really focus on visual discovery, in addition to the augmented reality overlays for which it was initially known.

Back to Google

Blippar offers a slightly different approach to Pinterest — providing a full camera-based browser that continually analyses anything you put in front of it. Once it recognises an object, you can access more information or take a number of other actions too.

Both of these examples are far more intuitive than using words in terms of how people naturally shop.

Now, some of you might be thinking “but doesn’t Google also have super cool image search tech already?” and you’d be right to ask that question.

Yes it does, and Google was in this sector early too! You can visit and search using pre-existing images. But as a company, Google has never really pushed this functionality. Somewhat bizarrely, the best advert for this ability has probably been the MTV show Catfish, (please, don’t judge me) where they seemingly use this every episode to search for online impersonators.

I suspect the reason Google hasn’t made a major push in this direction is one of three things (or a combination thereof).

  1. It hasn’t yet successfully monetised the results from image searches (see below) in the same way and doesn’t want to further (and potentially unnecessarily) complicate its advertising structure.
  2. All the data it has (and it has looooads) in terms of user behaviour using this visual search indicates that users are using this more as either a fact-finding or inspiration tool rather than as a precursor to buying something. People may also just literally be using it as a way to find images… The knock on impact of that being that if Google served ads, advertisers would likely get poor results as they wouldn’t be matched to the actual intent of the user.
  3. People are genuinely only using this ability to catch out online ‘catfish’

OK, probably not option 3… but the fact you don’t see ads in image searching is pretty notable when you’re so accustomed to seeing them in search results.

So — why am I going on about this?

Why is Nick going on about this?

There’s something interesting going on here in terms of a conjunction of a number of things.

  • The current state of search advertising with ‘inspirational’ rather than ‘purchase’ intent
  • The availability and quality of image recognition and machine learning technologies
  • Strong indications that people are adopting and comfortable with camera-based interfaces
  • A number of challengers (eg. not Google) choosing to take up the mantle of ‘visual discovery’
  • A (seemingly at least) as of yet, unmonetized method of online discovery
  • The fact that I (personally) clearly over think precisely how I discover new trainers

If we’re willing to accept a traditional marketing funnel as a model, you might draw the conclusion that visual search will forever be upper funnel “awareness” and “consideration” activity rather than the actual conversion point — but that’s still incredibly important in reaching potential customers.

It’s also relatively liberating in considering the type of advertising that might resonate with people compared with the conversion-driven messaging of PPC search ads. Without necessarily needing to push, or even offer a sale directly, there are some interesting possibilities in the types of formats advertisers can experiment with.

Moving one step further from messaging, you could argue even the very channels we think of as ‘discovery’ should expand from the search engines we traditionally think of. Pokemon Go has already provided a proof of concept that a mass appeal camera app that overlays information over the physical world can work. It is (present tense — yep some people still play it!) also providing a reason for lots of people to explore the world around them more than they may have previously. ‘Discovery’ in a very real sense.

Augmented Reality & Advertising

There’s actually already a lot of experimentation in this area with augmented reality and advertising too. Platforms are keen to be the first to market with a monetisation of visual discovery that provides tangible rewards. Snapchat anticipates an opportunity for advertisers to win brownie points by offering branded, filters / lenses based on what a user has included in a picture. The example they use ( article here) is offering a ‘King Kong’ filter when someone takes a photo of the Empire State Building.

Optional Interactive bit! (Dual screening required)

Another (probably more commercial) example might be the big coffee chains competing to offer a filter for any coffee related pictures. This kind of ‘soft’ brand touch offers interesting possibilities for engagement.

Blippar talked at WXG 2017 and addressed the extent to which Augmented Reality is shifting from being perceived as an innovation to something that can be utilised strategically. Here’s a quick example of how this can work:

  1. Download Blippar (Android / iOS)
  2. Open it
  3. Point your phone camera at the picture below

Fun, right?

This kind of AR project using static imagery has been around for quite a while now — ( we’ve built one at Kyan too) but it nicely demonstrates how in a real world context this same tech could provide real utility.

While not ‘discovery’ per se, Facebook seems to be taking a slightly different (but related) approach, using image recognition to help advertisers categorise people into interest groups ( article here). The example provided is if someone is regularly posting pictures of dogs on Instagram, that person can safely be included in a ‘dog lovers’ interest group.

Looking to the future

In my opinion, it stands to reason that over time, as mainstream search habits catch up with the technical possibilities the (potential) distinction in commercial intent between keyword and image based searches may diminish. More commercially driven messaging would become more impactful at this point, with Google seemingly already well-prepared.

An obvious solution for Google, would be serving Google Shopping ads. These are the image based ads users are already served above the text ads and ‘organic’ search results (see below).


Whether this was in a search engine format as it already employs, or as some kind of augmented reality overlay, it would simply be a more refined version of the technology Google already has access to. It would stand to reason, that if I’m searching for / scanning a particular product / object I’d likely be interested in the prices surrounding it, even if I’m not quite ready to buy yet. Pinterest has already announced it’s going to offer similar AI-powered ads using its ‘related pins’ feature.

I can imagine this might be quite compelling for value brands in particular. Imagine someone uses a picture of a black t-shirt or something similarly generic. It might be a powerful ad impression to appear alongside more expensive brands for what is visually seen as a very similar product.

Even an industry like Search, something Google pretty much owns and has defined in the majority of global markets is potentially subject to the possibilities that new technologies offer for improving processes and user experiences.

There has been a lot of debate surrounding the role that voice search will play in how businesses attempt to optimise their digital marketing, but in my mind visual discovery is an even bigger opportunity. I’d argue over the coming years visual discovery has a chance to become the biggest shift in search behaviour since people started initially using mobile devices to source information and shop online.

Interestingly for marketers, the form in which this type of online discovery will eventually be most effectively monetised is (again, seemingly at least) as of yet, undetermined. There will likely be rewards for creative thinking when it comes to visual search marketing, whether that be in terms of content or the very channels and formats (beyond simply search engines) that are considered ‘discovery’ channels.

In terms of tangible landmarks on the near horizon, Samsung has already announced that the new flagship S8 device will come with Pinterest’s visual tech embedded into its native camera app for users to discover related ideas from their photos.

Meanwhile, Blippar already provides the ability to scan the faces of over 70,000 well-known public figures (and find more information on that person), and will shortly be announcing the ability for anyone to upload their own face to their face recognition database.

The final thing to monitor will be retailers implementing this kind of visual tech within their own applications. John Lewis has already seen success using this kind of ‘find similar’ product functionality to its e-commerce offering.

The task facing marketers will be to continually monitor this trend and assess the possibilities that a visual approach to discovery might offer for your business. Younger audiences already favour YouTube ‘how to’ videos over written articles, and they also make up the majority of Instagram and Snapchat users. I’m speculating that over the coming years, this demographic will also drive the growth of visual discovery to becoming a popular way to ‘discover’ online.

Originally published at on May 17, 2017.