For the first time since television, we are adding a new layer of experience to marketing with immersive 360 images and videos, but the importance of visuals is not only key for social engagement. Image SEO is witnessing a re-emergence due to not only Google, but also Facebook and Amazon, leveraging image recognition in user journeys. Learn what smart brands have done to make use these new visual technologies and get practical tips on how you can join their ranks.
10. @jes_scholz
TEXT BASED KEYWORD
SEARCHES ARE
INEFFICIENT IN HELPING
FIND THE RIGHT PRODUCT
ONLINE
- Slyce Visual Search Technology & Market
Report
74%OF CONSUMERS SAY
@jes_scholzVisual marketing
38. 27%OF SEARCHES ARE FOR
GOOGLE IMAGES
27%
GOOGLE
IMAGES
59%
GOOGLE
4%
YOUTUBE
10%
OTHER
Distribution of US Searches 2016
@jes_scholzVisual marketing
52. Structured markup
A vocabulary of HTML tags that improve the way
your page is represented in SERPs.
Read more from Rob Bucci or Dr Pete.
@jes_scholzVisual marketing
83. Spark - A social
discovery (read
shopping) network.
@jes_scholz
84. INNOVATORS
FEB 2014 JUL 2014 OCT 2014 NOV 2015 JUN 2016 NOV 2016 FEB 2017 APR 2017 MAY 2017 JUN 2017 AUG 2017 SEP 2017
Amazon: The early innovator @jes_scholzVisual marketing
85. INNOVATORS EARLY ADOPTERS
FEB 2014 JUL 2014 OCT 2014 NOV 2015 JUN 2016 NOV 2016 FEB 2017 APR 2017 MAY 2017 JUN 2017 AUG 2017 SEP 2017
Pinterest: The market educator @jes_scholzVisual marketing
86. EARLY ADOPTERS GROWTH
FEB 2014 JUL 2014 OCT 2014 NOV 2015 JUN 2016 NOV 2016 FEB 2017 APR 2017 MAY 2017 JUN 2017 AUG 2017 SEP 2017
Visual search has crossed the chasm @jes_scholzVisual marketing
INNOVATORS
87. EARLY ADOPTERS GROWTH
FEB 2014 JUL 2014 OCT 2014 NOV 2015 JUN 2016 NOV 2016 FEB 2017 APR 2017 MAY 2017 JUN 2017 AUG 2017 SEP 2017
Battle of the brands to be your snap to act app @jes_scholzVisual marketing
INNOVATORS
98. Use case 2
Product identification for
snap to sell
@jes_scholzVisual marketing
99. Shoes
High heel
Pink Jimmy ChooOOOoo’s!!
6
£350
New without tags
Jimmy Choo
Pink
#$%@
Category
Subcategory
Title
Size
Price
Condition
Brand
Colour
Material
@jes_scholzVisual marketing
101. Shoes
High heel
Jimmy Choo Pink High Heels
6
£350
New without tags
Jimmy Choo
Pink
Suede
Category
Subcategory
Title
Size
Price
Condition
Brand
Colour
Material
Size
Price
Condition
@jes_scholzVisual marketing
I always love coming to the UK because the men are always dressed well.
From the 1900 suit, vest and tie combinations.
Through the style storm of the 70s
And on to full fashion force in the 2000s. UK men just look so darn dapper. And now of cause you have achieved a new pinnacle of style and sophistication. I speak of cause...
Of the checked shirt. Modelled here by fashion icon ed sheeran. I know all the gentlemen in this room face a common dilemma. You want to buy this exact shirt.
So You visit google and type “ed sheeran check shirt”.
You click through the results, but while you can find style advice and even some similar items, you can find that exact shirt for sale.
And So you turn to image search. But clicking images lands you on the same informational pages.
You try to qualify your search by adding by online and then everything goes downhill. You end up frustrated without a shirt, and there is no sale for the retailers.
74 percent of consumers say that text-based keyword searches are inefficient in helping to find the right product online.
This inefficiency of image search plagues not only ecommerce sites, but also content sites and even the service industry. It an equal opportunity fail.
But really is this any wonder why we had this inefficiency when a search engines take a text-based query and try to find the best visual match but image optimization, which should powers those recommendations.
is a big pink dinosaur in the SEO world. We all know it’s there, but it hasn’t historically contributed to KPIs so we just kind of ignore it whenever it is in the room.
EXAMPLE. The hard truth is SEO professionals often neglect image optimization as an afterthought to page-level optimization, if it’s considered at all. So why am I standing up here talking to you about it…
EXAMPLE. But because you really want to ran for the page level keywords, you call it pancake recipe ideas. Or maybe even
EXAMPLE.
The hard truth is SEO professionals often neglect image optimization as an afterthought to page-level optimization, if it’s considered at all. So why am I standing up here talking to you about it…
Because there has been a breakthrough in image recognition technology thanks to computer vision, machine learning and the ability to delve into metadata.
These technologies allow computers recognise the content of an image and based on that visual content return relevant information. There’s no need for keywords any more.
The functionality is called visual search.
Let’s take our scenario from before
So you see something you want to buy, like our checked shirt, so what do you do to find it? You search but now...
Unlike what happened in image search. With visual search I can take the picture of ed sheeren as the search query.
That one picture, it’s worth one thousand keywords. We are no longer reliant on SEOd tag to understand images. This also solves the problem of the user trying to describe the image, this eroded language barriers allowing visual content to engage a wider audience.
We live in a mobile first world. So it can’t be a surprise that much of the visual search functionality is only available on your phone. You need to get off your laptop at work in order to understand consumer behaviour and see how your brand is positioned against your competition. Otherwise you’re just doing a laptop selfie.
Because google has been quietly adding machine learning and image recognition capabilities to mobile image search over the last years. Plotting these updates, we can see clear stepping stone technologies building on the theme of visual search. Back in April 2013 Google released had related images functionality where you would click on a result to view visually similar images.
The collections update allows users to save images directly to folders. Google’s answer to a Pinterest board, which came into popularity around the same time
In 2016, product images begin to display next to website links in mobile search.
And later that year, clicking on an image result would display product price, availability, ratings, and other key information.
But things really got interesting in April, when Google launched to identify products, even within lifestyle images. Now we see we are able to shop the look of fashion icons.
But that’s not all. 3 days later Google released style ideas. The flip side to similar items. If I have for example found a pair of shoes I like, style ideas shows outfit montages and inspirational lifestyle photos to highlight how it can be worn in real life.
And finally, Google released labels on image that indicate what other details are available, encouraging more users to click. For example, the label recipe, or timestamp for pages featuring videos, or “product” – shown if the item is available for purchase online.
When you put all these updates together, you see powerful functionality. Google is making a play to turn image search into a product discovery tool.
And I wouldn’t be surprised if Google’s next step is to mimic Pinterest, who’s flashlight technologies also allows user to select to search objects in images.
Especially because Bing has already done that.
Tbecome a tool not only for search, but for discovery. Encouraging more people to begin their shopping journey on the search engine. This is a fundamental change in image search and with it a big opportunity for early adopters.
Discover through images is alive and well with 27% of all searches being for images. And this number is only going to grow as young consumers are increasingly visual.
What’s more impressive is that 11% of text based Google searches have an image feature box. Even the search image block has an image block.
And this is not only where you would expect it, on transactional queries, such as handbag London.
Images blocks are also showing up on informational search queries. Here we can see the search query “Kate Middleton style” render an image block. And when you click on the image result, you can shop her look. This richer user experience is educating searchers about Google’s shopping functionality.
To understand the impact this could have on my business, I turned to Moz SERP feature report. Here all I can see all the various SERP features but take special note of the coverage of image pack. Not only did I see that images were among the most common SERP features for on publishing websites for informational queries.
But also the most common on our ecommerce websites for transactional queries. And so I rolled out visual search optimisations to capitalise on the opportunity. And this was the result for one of my websites in Tanzania...
When you type in handbag tanzania we are dominating the search query taken the number 1 organic spot with an image in the web result.
As well as further down the page in the image page
And in the image search results we are marked up with product labels, unlike our competitors and if you click though, which you are more inclined to do thanks to the lable..
And full details when you select that product. So when people click from image search to my website, they are a high quality, validated users who is unlikely to bounce.
And this is not for only one search, but many where you can rank images. And it’s not limited to traditional ecommerce products, it also works on big ticket items, such as flats for rent and cars or informational queries, such as recipes.
And it doesn’t matter what the product is. I’ve been giving a lot of ecommerce examples, but we have this working not only for traditional ecommerce products, but also used cars and even rental property.
Here you can see the impact of rolling out image optimisations which can be up to a 25% lift in image impressions in only 30 days.
To compete is a world of search that is not driven by keywords, but images you need to stop prioritising webpage keyword level optimisation and invest time to optimise for image blocks to outrank your competitors in the SERPs. To do this, there are 5 easy steps.
Start by reviewing the the images on your site. Look at the quality. High res images where the product is clearly recognisable as being that product as will outperform those which do not. Don’t use images where the product is obscured or the scene is cluttered.
I don’t have enough time to go into all the details of structured markup, so for now know that it is a ….. For more info chat to Rob Bucci or read the articles by Dr Pete.
But essentially you want to use structured markup to tell the search engines as much about the product associated with that image as you can by implementing schema.org product markup. Paying special attention to ---- as these 5 features are what is going to earn you that label badge.
Validate your implementation in the structured data testing tool.
It can take up to 1 week for Google to index all your marked up images. So help it along by submitting an image XML sitemap. After a few days you should start seeing the labels appear on your images
Do a site query site: to see your images…. Now do you see a problem with this image? Yeh. It’s missing the markup. Now it’s true that Google doesn’t not guarantee that label just because you have the markup. But I can tell you that it will definitely not show right now if you are on your desktop. Remember...
No laptop selfies! You’re not going to be able to see this markup on your computer.
Do your checks on your mobile. So you can see what your users see. Doing these image optimisations will give you a quick win in the short term.
But all of this presumes the user sees something they want and then searches. The current paradigm for marketers is wait for a search and compete for attention.
But with Image recognition you can cut out the search step in the customer journey, and take the user directly from desire to conversion. Brand can enter the funnel before the search occurs, effectively excluding their competitors. Who cares if they have #1 organic spot, or better Adwords or effective messaging if customers don’t see it because the research stage is removed.
This is what I call snap to act, when you connect the moment of discovery directly to the conversion by using the camera as your search box. I want to show you how easy buying can be...
This see-snap-buy behaviour is becoming commonplace and opens up opportunities for companies to enter the funnel with the photo as the search query. Allowing people to spontaneously find anything they see offline, online.
Now I know this may seem reminiscent of QR codes. QR codes never really took off because people didn’t understand how they worked. But everyone knows how to use a camera and It doesn’t take too much to imagine how brands might be able to use this functionality in interesting and engaging ways
A billboard can now act like a banner ad, and deep link to a special offer. All of a sudden above the line ads are directly actionable and trackable. It’s like QR codes on steroids.
A magazine ad can now allow you to snap to shop each individual item.
Shop is closed for the day. Not a problem. Snap a pic of the item you want and have the clothes delivered to your house. Storefront become an asset 24 hours a day. This is transformative for brick and mortar stores because if you don’t offer your products via snap to shop, your competitors could show similar items available for sale when potential customer take photos of your store.
Even people become ads as they are walking past on the street. With this technology, you can literally buy the shirt of someone’s back.
We are seeing a merger of search engine, social networks and consumer brands all wanting to become your snap to act app of choice.
Google is launching Lens. Essentially the next generation of Google Goggles, it is powered by augmented reality drawing on visual search, to deliver actionable info based on what your camera sees. If you’re a local SEO’s get excited, this technology allows the simple act of scanning a business storefront for Google Local results to pop up, complete with reviews, menus and contact details. Allowing brick-and-mortar stores to capitalize on hyper-local searches.
You don’t need to have a product to sell to use the technology. For Commonwealth Bank of Australia mobile app with image recognition is changing the house hunting experience. A prospective buyer can take a photo of the house they like.
The app identifies the house and provides the list price, taxes and other details. The connects with the buyer’s personal financial information to give mortgage options.
The static world that was not digitally connected, now can now be digitlized simply by pointing your phone at it.
But search engines aren’t alone. Visual search has become a strategic advantage for Pinterest as they have positioned the platform as search engine for online discovery and leads the pack in terms of integrating visual search with shopping, and having a jolly good jab at Google while they are at it.
In direct competition with Google, Pinterest also has Lens (the name Google copies). Pinterest Lens helps you get online recommendations for similar items on product you discover offline. For example. if you point Lens at a pair of red sneakers, Lens will find you visually similar sneakers as well as ideas how to style it.
But its not just ecommerce websites that can leverage Lens. Content sites can as well. For example, if you take a photo of strawberries you won't get more strawberry photos. Pinterest understand you don’t want to see more pictures of strawberries, but you’re seeking information. So will show you recipe ideas.
And even brick and mortar stores are recognising the potential to improve user experience. In an exclusive partnership, Pinterest Lens will become available in the Target mobile app, to scan objects in the real world and be shown a list of similar items sold by Target.
But I suggest you move away from thinking native apps. There is a clear trend these are on the decline. If you see no choice but app, at least build in a progressive web app. But I suggest you think more on how you can utilise image recognition where you can get in the middle of an existing customer journey. Rather than trying to divert a customer's attention to you.
Already we see smart speakers, digital personal assistants and chatbots on messengeing platforms taking over from the myriad of mobile apps because they are simply more convenient.
The time of the website and the native app where the user have to navigate to you, search for you, is coming to an end. You need to go to them. And think about how you can distribute your content through bots, skills and other integrations. Which is why I think Pinterest is taking a smarter approach with their partnership with Bixby - samsung’s personal assistant.
Pinterest Lens powers Samsung’s Bixby visual search. When you point at an item and click images icon, you are directly linked to Pinterest results. It’s the first time the social media company has powered an in-build personal assistant on a mobile device, rather than being only a pre-installed yet standalone app.
But that’s not the only powerhouse partnership in Bixby. It’s shopping icon link to Amazon. But dont think that Pinterest and amazon get along just because they are both integrated in Bixby. They are direct rivals in terms of visual search technology with amazon it’s own in it’s native app.
And the rivalry doesn’t end there. Amazon also facies itself as a social network with Spark. A flow of pretty pictures with pretty products, where you can share your experiences to get smiles. With all images shoppable thanks to image recognition. Its essentially a big social network of product reviews.
Of cause amazon has a jump start because they released their visual search tech all the way back in 2014.
Followed closely by other ecommerce shops and of cause also Pinterest, who has taken a clear lead in market education on how search by images is more effective than keywords. During this time of market education, it went quiet. Until advances in image recognition caught up with user desire.
search engines joined in, Bing slightly later than Google
And the floodgates opened as we see that visual search has crossed the chasm. Brands are battling to become your snap to act app of choice.
This is why so many companies are excited about image recognition. Visual search resulted in 48% more product views. 75% more loyalty. 51% higher time on site and 9% more average order value.
So is it any wonder why we have ecommerce giants pretending to be social networks, social networks wanting to become search engines, search engines trying to be social networks and a whole bunch of other companies in between. And the only consistent in all this crazy is that every brand wants to master visual search to drive more online sales. And then there is you. How can your brand compete?
Stop focussing on keywords. Learn to prepare for when your camera becomes the primary search box.
It’s really quite simple. You just need to master image recognition. To do this code a query to image understanding...So that the basics. everyone here know how to do that right? Yeh, maybe not. For those of us who don’t know this stuff, you just need 1 simple formula
https://blogs.bing.com/search-quality-insights/2017-06/beyond-text-queries-searching-with-bing-visual-search
Developers + some beer over the course of 1 day hackathon. This is what you need to develop image recognition capabilities. It won’t be perfect, but thats OK because it’s not about releasing it to end users. It’s about getting an MVP to show your c-level to get buy-in in invest further in image recognition technologies. Because this is new. C-levels are not going to invest unless they understand the possibilities and they won’t understand the possibilities if you can’t show it to them.
I know you're asking how can my developers build this in just one day? There has been a increase in the number of services offering the ability to recognize an object, logo or person in your images. Many from big players like Google with cloud vision, IBM with Watson, Amazon with Rekognition and salesforce with einstein. but also smaller startup companies with slyce or clarifai. They can leverage pre-trained image classifiers or train your own custom classifier. This is how you can utilise the power of image recognition for your brand to drive conversion.
But now the question is, what do you want to build. You know your use case best, but here are some ideas to get you started.
These out of the box API solutions are pretty good at classification. At least, they know this dress is blue and black. More than it is white and gold. You connect the API so user can upload images and compare against your database to find similar items.
Or Thanks to facial recognition, I can snap a photo of a music artist and link to his songs for download.
But it’s not only for buying item. You can also use image recognition to better classify user generated items for sale to drive listing conversions.
I want to sell these shoes, you need certain information…. And I get to a point where I don’t want to your questions I just want to list my shoes for sale. So I either leave the conversion funnel frustrated, or skip all the non-compulsory fields. Either way you have poorer site content.
Why are we doing it this way when image recognition can understand this image
If we started with the photo. I know most of the fields. All I need to know from you is the specifics of this individual item. Save your questions for the answers that matter.
Ebay has only just added visual search doesn’t use image recognition during the listing process.
The startup LetGo does use image recognition and starts the listing process with the photo. This 2 year old brand has stolen 20M monthly active users form the incumbent.
The problem I see with LetGo is that they hinder their own growth because they require uers to download their native app to list. Rather than 20M they have, they could reach 1.2 Billion if they integrated with Facebook Messenger and let me list in a chatbot, taking the download the app step out completely of the conversion process by going to where the customer is.
What if you don’t buy or sell products. You can image recognition to provide information. Take a photo, get an answer is the most basic form of visual search. And it can answer a whole bunch of question such as how can I get this girl to date me...
It could also answer, what could I eat tonight. Fridges are having camera’s installed which link to my Google Home or Alexa, designed to order more food as I run out. But you can use that hardware. Now say you are a recipe website, you can create a skill which takes that information from the fridge and recommends what to eat for dinner. This is not only useful & personalised to the user, but future proofs your brand for when your native app and website go on the decline.
And local SEOs, I have not forgotten you. A restaurant could put up a bus stop ad which integrates with Google maps to show public transport directions to their restaurant. Take the customer journey quite literally. Tell them where to get off the bus.
So users have seen something they want, they can find it using visual search or snap to convert with their virtual assistant or in their app of choice, more likely than not, one run by Google or Pinterest, or some other large player. But visual marketing doesn’t stop there. It can help you post conversion as well to help you understand and drive customer advocacy.
Though something called visual listening.
We are all familiar with social listening. The opportunity to hear what customers think and feel about your brand by tracking conversations that mention specific words or hashtags. like brand name. The problem is, if you’re only paying attention to text mentions, you’re missing most of the conversation.
Around 3 billion images shared each day, but over 85% of social media posts don’t contain any tag or text brand mention. Text based insights alone wont give the full picture. This why you need visual listening - based on logo recognition it is a more targeted use of image recognition - it can uncover all those in image mentions of your brand and drive a stronger user understanding.
This tweet includes the Coca-Cola logo and an interesting looking hamburger, but it doesn’t reference Coca-Cola in the tweet. Without visual listening, Coke would miss this mention. Actually without visual listing...
But how insightful is a hamburger? It’s not going to be a shock to coke that people eat fast food and drink coke. What can we get that’s useful?
If nothing else, with visual listening you have access to 80% more text from the posts that didn’t mention your brand for sentiment analysis. Which here you don’t need to be able to read the text to tell the sentiment is no good for coke.
You can get insights into consumer behaviour and use cases to find out how your customers really use your product. For example, did nestle know those who are health conscious amongst us may eat their cereal with yogurt rather than milk.
Or insights into purchase behavior and what consumers really think of your prices. Because apparently Intel’s prices are meme worthy.
You can also understand what other brand are often visually associated with yours, to build partnership & reach a wider audience together. Or if your already have partnerships in place, how do your customer respond.
You can find those ever so precious product advocates and influencers. Because I tell you, if Adidas has logo recognition, they are very happy right now. Selena gomez is promoting them for free as a natural brand advocate.
But most importantly, you can find gold. Visual user generated content. Image recognition can allow you to find not only logos in images, but developing to a level where it can identify the products themselves.
And this is important because 86% of us UGC is a good indicator of brand quality.
Use logo recognition and product identification to find advocates. Find UGC in that 80% of social mentions that are not taggig your brand
And once you find it, with the user permission, showcase your community’s photos on your website, preferably direct on the product page, as well as in newsletter and on your social media channels.
Use visual search to make the products in the pics purchasable. So user generated content acts as a visual customer review and assist to drive conversions.
… to get your started…. Are my top 5.
Turn selfies into sales to have a truly formidable marketing strategy. Save on costs of content creation while you build brand advocates by finding authentic UGC with logo recognition to significantly increase brand engagement.
With UGC driving product advocacy, it will feed back into additional discovery opportunities and you have a self-sustaining marketing model driven by visual marketing technologies. This is the power of visual marketing.
If you take 1 thing away from this talk, it needs to be these 4 words.