Kookai : Okkular Tag-Gen solution Case Study

Ecommerce data standardization is essential for optimizing business processes. Businesses across the globe depend on this product data to increase their understanding of consumer behaviour to make qualitative decisions. Therefore, the data must be consistent, regularly updated and easy to manage. Inconsistent product data can slow down the digitization of products and the time it takes to get a product to the shopper.

Okkular’s state-of-the-art Tag-Generation capabilities can help solve these problems for retailers. Our deeper attribute-related information generates product tags and metadata that are SEO/SEM friendly within seconds. We got in touch with Kookai, a leading fashion retail brand to understand how Okkular has helped them in this journey.

About Kookai:

KOOKAÏ is an Australian owned women’s fashion label, defined by an essence of femininity, confidence, allure and independent style. The KOOKAÏ collection is designed exclusively out of our Melbourne and Paris studios through a creative, symbiotic process forever inspired by the KOOKAÏ woman and, in turn, create pieces through which to continually inspire her.

  1. What business problem were you trying to solve? 

We were looking for an efficient way to enrich product data to avoid cart abandonment, provide rich meta-tags, accurate product descriptions and categorizations. Given our high volume of SKUs, we needed a product that could assist with optimising & improving our website search functionality as well as contributing to improving our SEO/SEM. 

2. How was the problem affecting your business?

We were in the process of implementing a new website search solution which required updating our product catalogue using tags to assist in enriching search data and filters. This would have required manual data entry which is cost, labour and time intensive. Automating this function would free up time, cost and help implement changes in much smoother and faster way.

3. What options did you consider to solve this problem?

Other visual search solutions were considered however Okkular’s Tag-Generator stood out as we were able to easily identify how that would immediately benefit our business due to our wide range of SKUs. We were looking for a product that can help ingest large data and process its efficacy.

4. What reservations did you have before choosing Okkular?

  • The accuracy of the AI data given our wide range of apparel and accessories
  • Aligning Okkular’s AI naming conventions with our brand’s style classifications
  • The ability to easily integrate and automate sending the analysed data back to Shopify to work cohesively with our search solution.

5. Why did you choose Okkular?

Okkular’s auto Tag-Generator is extremely beneficial given our high volume of SKUs. The automation of Okkular’s AI enriching our product eliminated the process of manually generating tags which would have resulted in hours of manual work. In addition to this, we were excited to work with a Melbourne start-up tech company who have continuously provided ongoing support.  

6. What are some of the immediate business benefits you have noticed after implementing Okkular Tag-Gen solution?

  • Reduction in manual data entry of tagging product
  • Improvement of search results
  • Improved filter options
  • Keyword friendly tag
  • Better on-site navigation

7. What are the long-term benefits you see automating the Tag-Gen process?

As Okkular’s AI continues to learn and analyse our evolving range, the time has been reduced in auditing and verifying tags within the Okkular portal due to the result of more accurately generated tags. Additionally, we are working with Okkular to begin using the automated alt-text feature to assist in boosting SEO/SEM.

8. What industry trends drove the need to use our product?

We wanted to tick all the boxes for a good customer experience, catch up to industry and eCommerce best practices.  As mentioned previously, we were optimizing the site for the front-end consumer by giving them an improved shopping experience and wanted a succinct search solution and improved functionality.

9. What feature of the product was most appealing to you?

Okkular’s Tag-Gen matched most of our requirements, especially from the point of view of improving product catalogue data and user experience.  We particularly, wanted something that could work cohesively with Shopify and Okkular’s Tag-Gen has those parts in place. From the standpoint of overall improved search functionally, Okkular’s Tag-Gen has provided KooKai with the kind support it needs for a holistic online shopping experience.

SEO hacks for E-Commerce

Online visibility today is no more a luxury but a necessity. Over the years, the number of e-commerce operators has greatly increased and it will continue to do so. In this post COVID world, an online presence is everything. So, in this sea of online retailer how can you make sure that your website attracts maximum traffic, ORGANICALLY? 

Let’s look at a few things you can implement to drive traffic, conversions and improve your Google ranking.  

Product Tagging

Tagging is the process of assigning keywords and phrases to a product so that it’s easier to locate on your e-commerce website.    

Typically, products are assigned to categories and are searchable via keywords. Not only do the tags highlight the attributes of an item, but you can also use these tags for cross-referencing other similar products. When a customer is viewing one product, they could click on one of these tags to see a list of other products with the same tag.  

It is no secret that tags improve a websites on-site search and has the highest rate of conversion for e-commerce product pages. Properly allocated tag for a product can also improve your search engine rankings as they double up as keywords. In some cases, this may trump even brand-specific search engine queries. Therefore, intuitive product tagging is a must for online retailer/e-commerce operators.

To a search engine, these tags are full of keyword-rich links to relevant content deeper in a site. As mentioned before, each piece of tagged content should have links so the users can find other content related to it by tags, and search engines really like this. Tags are great for SEO too because your tags generate their own URLs, and each tag is a keyword-rich internal link to that page, reinforced by the tagged items themselves and the sitewide tag cloud if you have one 

At Okkular, or AI-powered Tag-Gen can ingest any online catalogue and produce intuitive, attribute specific and SEO friendly tags, guaranteed to improve your on-site search as well as search engine ranking. The labour-intensive and time-consuming process of manually adding tags can now be handled in a matter of seconds with our AI Tag Gen. Dynamic and customizable as per your needs, the Tag-Gen is constantly learning, developing and absorbing new tags keeping in mind the consumers search habits and latest trends.     

Note: Make sure when you design, your category and product pages you use tags that are interlinked and keyword-rich. The objective is to hold the interest of the consumer by making navigation easy and optimum for the final conversion i.e., purchase of the product.   

Let’s now deep dive a little more conceptually to understand more about different kinds of SEO keywords and strategies. 

LSI Keywords

Back in the day, if you wanted your page to rank cramming it with one keyword repeatedly was your best bet. Google and other search engines would figure out a page’s topic-based 100% on the keywords they found on the page, which is why keyword density was very important.

However, things have changed since then, the new google search algorithm ‘Panda’ and ‘Hummingbird’ have greatly changed the way the crawler indexes and ranks pages. In simpler words, it now uses ‘context’ to determine the credibility of a page.    

Latent Semantic Indexing or LSI keywords are conceptually related terms that search engines use to deeply understand the content on a webpage. Google relies on LSI keywords to understand content at such a deep level.

For example, let’s say you added a new product page for a floral-print dress. Google will still scan your page to see if you use the term “floral-print dress” in your title tag, content, image alt text, etc. At the same time, Google will also scan your page for LSI keywords like “autumn, “fall collection”, “midi”, “print”, and “dress length”. 

And when they see these LSI keywords in your content, they’ll say: “We’re confident that this page is about a floral print dress.” 

Keyword Synonyms

The mind of a consumer works in myriad ways, each consumer will have different ways of identifying products and searching for products. A single product can be identified using different keywords depending on the location, language and preference of a consumer.   

For example, “Oxford-Shoes” are also called ‘Bal-type” a user can use both the terms to search for the same product. In France Oxford shoes are called ‘Richelieu’ so a user in France or of French-origin might use the word ‘Richelieu’ to search for the same product.  

Let’s take another example, a ‘blouse’ is also known as a ‘top’, a consumer may use two different words to look for the same product. However, the term blouse in India would identify with clothing that is worn with a Saree. It is important to understand your audience and optimize your page accordingly.  

Integrating your product with the relevant synonyms help combine multiple searches and direct the consumers to the relevant page. These simple tweaks will ensure greater conversion and help improve overall search rankings.

However, identifying these keywords can be tricky, for such SEO optimization you’ll need a deeper understanding of each product, their synonyms and search volumes. A fashion enthusiast combined with an SEO expert would be the ideal fit for such an optimization. Unfortunately, often fashion enthusiasts love fashion and not SEO. 

At Okkular we help users find the perfect keyword-rich tags for the product effortlessly. Our highly intuitive tag-gen can identify related tags and synonyms that can be optimized for high SEO traffic and visibility.    

Long-tail Keywords

Consumer search habits are diverse, each has its own way of looking up products. Some will your synonyms, some may use a specific keyword, there is another brand of consumer who may use multiple keywords to search for a product. Often consumer may string together a slurry of keywords to look for a very specific product, these are called longtail keywords.   

Let’s look at an example, a consumer may want to buy a wedding dress, his/her search query may look something like this ‘lace-mermaid-eggshell-wedding-gown’. The query is very specific, the consumer wants a mermaid style wedding with lace and in eggshell white. The keyword is attributed to heavy and very specific, and the consumer is more motivated to buy this item if he/she finds it. 

How do you then as an e-commerce seller capture this as part of your SEO process?  

Here is how we can help – our AI closely examines the product and produces ‘tags’ that are attribute specific with each detail of the product listed down. Our tags help push your product in more complex searches, and  give  you  access  to  another  segment  of users. 

These attribute specific tags not only help optimize search engine ranks but also on-site searches as well. You can customize your tags as per occasion, events, trends within a few minutes. For instance, Valentine’s day is around the corner and you’ve noticed a jump in the search for red dresses. With the help of our Tag-Gen, you can select all the dresses marked red add the prefix “Valentine’ or Valentine’s Day’ to it and optimize your website in just a few clicks, without the pain of manually doing it for each product. Save on time, money and effort and drive the trending traffic to your platform with ease of our products.  

Long-tail keyword and Speaking URLs

Now, this is hardly news that e-commerce platforms are riddled with the problem of duplication. When you have products that are similar, often creating unique URLs for them is a pain point for web-developers and SEO managers. You end up creating similar URLs that confuses the google crawler and causes SEO loss. Often one URL ends up cannibalizing the other, diminishing its traffic completely. This results in one item getting sold out completely and the other piling up in your inventory.

Our system generated attribute specific tags can solve this problem for you. Here’s how? Once a catalogue is ingested our the Tag-Gen will start producing keywords accommodating for each aspect of your product, based on its length, brand, dress type, neckline etc.  

A typical tag generation would look like this:   

body fit; skin tone; knee-length; round neck; dress,   

Not only do the tags generated resemble a long tail keyword it resembles a ‘Speaking URL’. What is a speaking URL you may ask? These are SEO friendly URLs that help generate more traffic and help increase your search engine discoverability.   

A typical URL looks like this: 


This is what a speaking URL looks like: 


An easy to read URL with a high discoverability rate, enriched with the goodness of a long-tail keyword. Creating these URLs is fairly simple with the help of our Tag-Gen, feed your product to our Tag-Gen, acquire custom tags for the product and create a unique URL for your product without the problem of duplication. 

SEO is a holistic effort of all pieces of a business, including social media, marketing, web design, networking and copywriting. However, when you optimize your site, you create the best possible result for your target keyword. 

The common pain points of an e-commerce platform stem from the abandonment of shopping cart, unavailability of the item or simply deciding that particular product wasn’t exactly what they wanted. Okkular’s Tag-Gen creates product tags that can help fill in these gaps and mitigate the losses these problems cause. 

And so, after you’ve made it this far, I must ask – are you really harnessing the full potential of your Product-Tags? 

What lies ahead? The Future of Fashion Retail

Since its outbreak in China last year December 2019, COVID-19 has sent shockwaves down the global economy, one of the worst-hit -fashion retail. Retailers globally shut down their doors for a whole season as the global pandemic took over our lives.  

Most fashion and luxury execs are convinced that the impact of the crisis will last for at least 12 months. A market that depends on offline sales, the impact of continued shelter-in-place orders and store closures will be tremendous. 

A BCG analysis conducted in March revealed a 60-70% decline in sales from April to May in the worldwide fashion and luxury industry – with footfall in retail and recreation stores down by 44% in the U.S., 52% in Germany, 78% in India, and 59% in Brazil.16. The remainder of 2020 can expect at least a 28-38% drop in sales of fashion and luxury brands. 

With June fast approaching, while things remain unclear economies across the globe are slowly opening up. The governments are relaxing the lock-down and retailers are getting ready for business. Despite this, the demand for fashion is yet to pick up as consumers remain focused on essential items.  

Speculations about the future of fashion retail are rife as the struggle to survive the coming months continues. Let’s look at what the data and consumer trends suggest:   

Consumer behaviour and sentiment:  

Consumers around the world are constrained to their homes. At least for the coming few months groceries, medicines and staples purchase will take priority over fashion and luxury items. And while online retail has sustained the crash, it is doing marginally better than offline retail.  

Despite this consumer sentiment towards fashion remains positive, though their behaviour is likely to change. Online retail will see a sharp incline as opposed to the highly marketable ‘shopping experience’ of retail stores. Brands will have to figure out ways to make online shopping a more wholesome experience.  

Early signs show that demand for goods associated with well-being, and the collective goodwill will enjoy more popularity. The consumer is more likely to spend consciously, think about durability and quality given expected economic uncertainty and high levels of unemployment among consumers.  

Sustainable Fashion:  

Macro trends indicate a demand significant action for sustainability fashion retail. Needless to say, brands will have implement changes in their procurement method. Inventory management will have to more conscious and sustainable given that climate change and many other environmental factors will drive changes in the fashion industry. And expectations for the entire industry will be reset around greater collaboration, more equitable partnerships, and collective responsibility across all parts of the value chain.  

It is the time to see how agile a company is in reacting to new market demands. And being a sustainable manufacturer with sustainable products should still be a strategic advantage. 

Credit: Weaving-a-Better-Future-Covid-19-BCG-SAC-Higg-Co-Report

Re-defining Retail:  

Fashion retailers are the ones bearing the maximum brunt of this unprecedented lockdown, without consumption, it is impossible to kickstart global supply chains.  

Experts across the globe are convinced that customer preferences may change, fashion retail will have to go through a transformation, a re-invention of biblical proportions of sort. The consumer will likely avoid physical outlets for at least the next 6-7 months. Going digital is now a prerogative for every brand out their if it wants to stay put in the market.   

In the 2003 SARS outbreak, the population hesitated to venture outdoors, that drove lasting changes in retail consumption patterns, most of the population shifted to online shopping. Until things normalized every brand must work on having a strong online presence and find the right ways of connecting with their customers. 

Defend business and its critical assets. Protect human capital, financial capital, and supply chain relationships. Manage orders while preserving supply chain relationships. Avoid cancelling completed orders to prevent factory closures and impacts to workers. Engage in collaborative conversations and respectful dialogue with suppliers on how to address issues together. 

It is important to note that retailers will have to find a liable solution to these issues as the frequency at which pandemics breakout has increased. Disruption to the supply chain can be debilitating to the business which is why retailers must not just explore online avenues but other novel ways to connect to it consumer. AI will have a significant role to play in the coming few months with companies offering state-of-the-art innovations to manage inventory, enhance product discovery, deep tagging – brands that can adapt faster will sustain this economic shock and move on to build better models.  

Unfortunately, most players may not survive this period of low demand. Pressures on the supply base will lead to many closures and a decline in supplier capabilities over the coming months.  

Scale will be a determining factor, smaller players without, a clear proposition, established consumer equity and a healthy balance sheet will face bankruptcy. Mature sizable companies will have to protect resources to weather this crisis. 

About Us:  

Okkular is a Melbourne based Australian tech startup. Okkular’s vision is harnessing deep learning driven artificial intelligence to help retailers improve customer experience, reduce costs and drive sales with insightful automation and personalization. Our solutions have been developed by having conversations with multiple fashion retailers.  

If you’d like to know more, log on to our website www.okkular.io 

Fashion trends for the ‘New-Normal’

The sprits of fashionistas across the globe cannot be doused by this global health crisis, a little deferred, maybe. When the designers started working on their spring 2020 collections, they were blindsided by the global health crisis. 

When the deliveries hit the stores, the whole world was cooped up inside trying to make sense of the situation – include brands across the globe.  Remote access to everything became the new global reality.  

However, here are a couple of snug 2020 fashion trends you can try right from the luxury of your home. And who knows maybe it’ll inspire you to shake off those COVID blues as well. 

Try these styles that are perfect for the season settling right in front of your eyes or rather through the window in most cases.

Lounge Ready bralettes  

Katie Holmes Khaite cashmere bralette and cardigan last summer was the talk of the town. The ‘It bra’ as we all know was sure to be a hit again come spring. It makes a luxurious case for hanging out at home in nothing other than your bra and undies. Ideal for the days you don’t feel like getting dressed or when the day doesn’t keep you busy with your deadlines. A perfect outfit to unwind, relax and do-nothing in.  

Put a spin of that leisure wear 

Dips, dyes, and swirls have landed on just about anything these days. Steps of the mundane color pallets of the leisure and spice it up with some bright swirls and swivels. Spend the day at home in a floaty tie-dye house dress or cosy up on the couch in a pair of dyed slipper socks. 

Finally, those Coordinated Cardigans 

Knitwear is a big hit these days is an understatement, it has dominated the 2019 and 2020 collections across the industry and is safe to say is here to stay. Bring this trend to your WFH wardrobe, try matching cardigan and top. You could coordinate it with your T-shirt or a dainty tank. It’s a nice, it’s comfortable and makes you look super neat with very little effort. Plus, it pairs well on top of both your pyjama-pants a well as your jeans. 

Try out these simple tricks to set your wardrobe apart, let some color and comfort seep into your life while we all make sense of the new normal.  

About Us: 

Okkular is a Melbourne based Australian tech startup. Okkular’s vision is harnessing Deep learning driven artificial intelligence to help retailers improve customer experience, reduce costs and drive sales with insightful automation and personalization. Our solutions have been developed by having conversations with multiple fashion retailers.  

If you’d like to know more, log on to our website www.okkular.io  

Get ready for Summers top trends.

Get ready for Summers top trends.

(image credit: Aje)

Summer is days away and there is an optimistic energy in womenswear this season.  The key trends your customers will be searching for online include silhouettes that emphasise drama and volume; think statement shoulders, deep V necklines, rising hemlines and mini dresses. ‘Romance’ and ‘Safari’ looks are also strong trends this season. Read below for our word tag suggestions!


Make sure your key word tags include the one-shoulder dress to pick up on this seasons hottest trend! This Summer it’s all about the shoulders, either bearing them with off the shoulder styles, or decorating them with puff sleeves or bow detailing. 

Tops and dresses with shoulder ‘puffs’ or ‘blouson sleeves’ are growing in popularity, with Google Trends experiencing search volume growth for ‘puff shoulder tops’. Ensure these key words are attributed to any items featuring these details.


Romantic looks continue to appeal in Womenswear, with many retailers giving us beautiful feminine blouses in light and textured fabrics, with blouson sleeves, shirring details, in florals and paisley prints. As customers become familiar with this look, be sure to tag words such as ‘sheer’ to capture those looking for these feminine styles in light floaty fabrications. 

The ‘bias satin skirt’ continues it’s popularity for another season, updated in this Summer’s hottest colours and prints.

Driven by trends for whimsy and fantasy in occasion wear, ‘fabric bow’ detailing is up-trending as a shoulder detail or cinching in waists. Feature bow details appeared on the recent global catwalks and on the red carpet at the 2019 Emmy awards, further confirming the trend and adding to it’s retail appeal. Tie your look together with a statement bow!

(image credit: Zimmermann)


The ‘Linen Blazer’ and ‘Safari Jacket’ bring us this Summer’s relaxed workwear items. The ‘safari trend’ is a nice update for the customer whose style is more minimalist and pared back. 

The elegance of a Safari Jacket in crisp neutrals offers a chic workwear option, with details such as multiple front pockets and self-fabric waist ties. Paired with ‘paper-bag waist’ pants, skirt or short, these items create a relaxed and sophisticated look. 

Several Australian brands offer a fresh take on shirting, with cropped shirts, wrap styles in crisp cottons, and volume in the shoulder bring refreshed aesthetics to the basic work shirt.

Relaxed suiting in a soft palette of cream, olive green, soft pink, and terracotta in crisp fabrications is providing a refreshing option for workwear styles.

(Image credit Aje Instagram)

#MarigoldYellow #Pastels #Neutrals

Worn head to toe, yellow continues to trend in saturated hues. Customers may search using a range of yellow terms including ‘golden’ ‘marigold’ or ‘mango’ so ensure these tags are captured to help boost search results and sales.

Neutral and earthy palettes offer customers a harmonious palette in line with the relaxed linen styles. Earthy colours continue to trend with palettes featuring terracotta, cream, and shades of green. Layer them over one another for a pared back approach.

Customers will also be searching for pastel hues this summer, including blush pink, sky blue, and lemon.  Hot pink also looks fresher this season when it’s paired with red. 

#AnimalPrint #Spots #Paisley

The popularity of ‘animal prints’ continue this Summer. As this print direction gains traction and customers become more distinctive in their preferred animal print, we suggest including additional tags such as ‘zebra’ or ‘snake’ to your products.

Other print directions include palm prints and paisley prints which work back with this seasons overall ‘safari vibe’. Spots remain a strong pattern trend, as well as gingham which works especially well for resort wear and swimwear.  

Our top recommended word tags for Womens Summer apparel:

  • ‘One-shoulder’ dress or tops
  • ‘Puff’ sleeves 
  • ‘Safari’ suit or jacket
  • ‘Paper bag waist’
  • ‘Yellow’, ‘Mango’, ‘Marigold’
  • ‘Spot’ and ‘Gingham’ print
  • ‘Animal print’ ‘Snakeskin’ ‘Zebra print’

Fossick Trends is a marketplace where Designers can search and shop for on demand ready-made trend reports. Downloadable PDF’s covering Womenswear, Menswear, Kidswear, and Accessories.

For more information regarding Fossick Trends, Check out https://www.fossicktrends.com

Contact Lyndal Wallis: lyndal@fossicktrends.com

10 types of content to increase eCommerce engagement

The way people shop for products has become a mishmash pathway, toggling between online and offline sources.  The customer journey is far from linear, which forces brands to be just as nimble in engaging with shoppers. It’s up to brands and marketplaces to create a variety of touch points and valuable content streams to encourage engagement, connection and loyalty.  

When marketers put themselves in their customers shoes and imagine their typical movements and actions, they can shift communications from being one-sided to conversational.  As brands develop campaigns with the intent to connect and engage, shopper interest increases.  

Sam enjoys the search of finding fresh outfits, not so much the shopping part as he prefers it to be timely and convenient. He follows his favourite brands on Instagram and Facebook and Sam is on several mailing lists due to previous purchases, receiving regular updates on new products and sales.  A typical day in Sam’s life, he wakes up early scrolls social media and scans emails, heads to the gym for an hour’s session, back home for breakfast and commutes to work.

After dinner one evening, Sam starts to wind down, falling onto the lounge for a few episodes on Netflix.  He has a quick thought about getting a new tank for the gym sometime this week, but being time poor with work he’s not sure if he’ll get to the shops.  Sam opens Chrome, a quick search for ‘men’s black tank gym.’ A few brands he knows show up, but there’s a new one he hadn’t seen yet – Brand X. A quick browse on site and he’s enjoying the styles, graphics and aesthetic of the website experience.  Without much thought, Sam is happy to pay $35 for a new gym tank opting to use Apple Pay, which makes things simple and easy.

As he’s waking up one morning, there’s a new email from Brand X, Sam glimpses new product drops and notices an article titled ‘The ultimate workout routine for men’.  Immediately capturing Sams attention, he already has gym on his mind and continues to read the article getting some awesome tips on improving his current routine.

man shopping online deliveries

That particular touch point is invaluable to building loyalty with Sam.  The brand is communicating with the intent of understanding the lifestyle of their consumers. For Sam, he feels like there’s more to be gained by opening Brand X emails with interesting articles aimed at helping him improve and excel in hobbies and interests he’s most passionate about.  

We’ve compiled 10 types of content for eCommerce businesses to implement that will not only increase engagement, but optimise organic reach across search engines.

  1. Unique product descriptions
  2. Image meta tags
  3. Website on page descriptions
  4. Social Media – image & text
  5. EDM’s – product updates sales, styling tips, news relevant to audience
  6. Customer reviews
  7. Youtube Videos
  8. Blog posts and interesting lifestyle articles
  9. Online communities
  10. Personalised thank-you notes

When arranging Sam’s day in a timeline of events, along with a loose customer experience map, brands can start to see how each touch point links to another.  The output is consistent brand messaging, enjoyable imagery and above all meaningful content that has a long-lasting resonance.  

Okkular Tag Gen uses image recognition technology to automatically attach tags to products based on patterns, colours and styles, making it easier for customers to search for the items they’re looking for.  Contact their team for a complimentary demonstration of Tag Gen.

Gamified Fashion Marketplace – Dresswhere.com, an Okkular Tag Gen Case Study


Key takeaways for using Okkular AI based Tag-gen solution

  • Over 8 weeks of full-time hours saved for tagging over 20k products. This helped in reducing cost and time significantly and go to market faster.
  • Produce over 20 product specific Tags/attributes for each apparel, which helped in making the search engine more efficient and robust
  • Over 80% accuracy on an average with AI classification and the ability to audit to make it 100% accurate 
  • Classify multiple colours which would have taken additional time and cost to add for every product manually. 

Intro to Dresswhere.com

Dresswhere is an upcoming aesthetic gamified search-engine for fashion. They don’t expect users to know what they want, so instead of asking them a very complex question (‘what are you looking for?’), they ask them a very simple question repeatedly: ‘which of these three apparels is best visually?’ Every time the user chooses an apparel or skips, they’re shown another three apparels that are slightly more specific to their tastes based on their previous choices. The effect is that, over the course of ~3 minutes, the user zooms in on an individually curated collection of apparels. Whenever they scroll down, they see their top results, and can click through to the retailer.

What was the pain point for Dresswhere to make the search engine efficient?

To create the Dresswhere search engine and predict a closer matching apparel, Dresswhere had to Tag every apparel’s image with at least 6-8 attributes/Tags. The catalogue inventory was over 20K products and requiring ongoing 3-4k products added every month. Dresswhere gets the product catalogue feed from multiple Fashion brands and the data received in the feed was inconsistent and was not enough for the search engine to work efficiently. While performing the task of tagging products manually, Dresswhere had two main issues.

  • Time taken/cost incurred to tag over 20k images and ongoing 3-4k products every month. Dresswhere estimated this would take approximately 9-10 weeks to tag over 20k products. 
  • The inconsistency and error rates

The above two issues were two key pain points to overcome for Dresswhere to successfully launch their product.

How did Okkular solve Dresswhere pain points?

Okkular.io has developed an AI based visual analysis solution which helps in automating the product tagging process. The AI tool generates tags from just the product images in under 15secs per product, increasing the speed to over 15 times faster compared with manual tagging. The tagging process also helps in reducing errors, generating consistent data and significantly reducing costs. 

As an example, Okkular Tag-gen solution can generate over 12 attributes for a dress including key attributes like neckline, style, length, print and also tags for occasions and avatars. The tags generated can be used in many ways including, titles, description, filters, product specific keywords, and meta tags.

Dresswhere provided us the product feed in a google feed format and the output of Tags were uploaded by connecting via Okkular API to Dresswhere platform

Working closely with Angas (Founder Dresswhere), we were able to develop additional key tags which were very specific to Dresswhere. We also trained and generated over 20 tags instead of 6 to 8 which was an initial requirement for Dresswhere if the task had to be performed manually.

During this project we frequently received feedback from Dresswhere to improve the classification accuracy, user interface and additional features which helped us co-create a solution which has not only helped Dresswhere but many other retailers using our solution currently.

What are the benefits for Dresswhere? 

  • Over 8 weeks of full-time hours saved. This helped in reducing cost and time significantly and go to market faster.
  • Produce over 20 Tags/attributes for each apparel, which helped in making the search engine more efficient and robust
  • 80% accuracy on an average with AI classification and the ability to audit to make it 100% accurate 
  • Classify colours from Apparel which would have taken additional time and cost to add for every product manually. 

Okkular is also working with Dresswhere to provide Okkular Visual search solution to help their users find similar looking products from multiple brands once they have identified a dress, they want using the Dresswhere search engine.  

If you would like to learn more about Okkular solutions, contact 

Mahendra Harish:  mharish@okkular.io 

For more information regarding Dresswhere, Check them out at https://www.dresswhere.com

Contact Angas Tiernan: angas@dresswhere.com 

Tips to help you improve product catalogue search and discoverability

Imagine this, Erin has a slight obsession with sunglasses. She already owns several pairs in various colours of similar style and shape because Erin knows what suits her face. She’s on a mission to find a chic pair of tortoiseshell sunglasses online and in a few clicks she’s landed on her favourite e-tailer site. Erin begins her search by firstly heading to the accessories menu, then into sunglasses. The excitement of an impending purchase is building.

That’s an impressive selection, 25 pages of sunglasses. Time is of the essence, she pans to the menu on the left to filter by colour and style. To her dismay, Erin can only filter colour and this leaves her with 400 products. This is not ideal as she knows exactly what she likes and scrolling through 12 pages is overwhelming and time consuming. The purchase inducing enthusiasm subsides quickly. Erin drops off the site and tries a long-tail keyword search instead on Google, which places her in prime position to find exactly what she’s looking for. She’s quickly matched with retailers who have their site optimised with clean content and relevant tags sorted.

94% of shoppers abandon a website or purchase when they are unable to find what they are looking for.

Generating product specific Tags creates a solid foundation of product data leading to benefits such as: improved on-site search, and search engine optimisation. Developing and maintaining product specific keywords results in a better user experience and increased conversion rate.     

88% of shoppers say that detailed product content is extremely important to their purchasing decisions.

We break things down into three tips, so you can get started on improving your product catalogue. Ensuring shoppers like Erin find what they’re looking for and keep that purchase excitement going.

Tip 1:  Amplify your content for better SEO and on-site search by including targeted keywords in product descriptions and relevant words in image meta-tags. This means shoppers using long-tail keywords in search queries will be more likely to see your products listed first.

Tip 2: Familiarise yourself with Google Trends and local search queries. Having a finger on the pulse in terms of real-time analytics of what shoppers are looking for, helps you to be dynamic in providing relevant product recommendations . This will also aid in keyword research, as trending products can change quickly and you’ll need to stay ahead of your competitors.

Did you know?
Manual tagging can take up to 30 hours for just 500 products

Tip 3: Switch to automated product tagging and create consistent product tags. This will cut down the time it takes to do the process manually and ensures all the words used to tag hundreds, even thousands of products is consistent. Tagging products manually is prone to mistakes, leading to incorrect data and a poor user experience.

50% of data workers waste their time finding and correcting errors or attempting to confirm data sources they do not trust.” Harvard Business Review

Okkular.io has developed an AI based visual analysis solution which helps is automating the product tagging process. The AI tool generates tags for your products in under 15secs/product increasing the speed to over 15 times compared with manual tagging. The tagging process also helps in reducing errors, generating consistent data and significantly reducing costs. 

As an example, the solution can generate over 12 attributes for a dress including key attributes like neckline, style, length, print and also tags for occasions and avatars. The tags generated can be used in many ways including, titles, description, filters, product specific keywords, and meta tags.

Okkular solution is simple to use with these easy steps.

Step 1:  Upload your product images or Provide us your product feed. Get quick results.

Step 2:Review and Validate the AI generated tags

Step 3:  Download ready to use SEO-Friendly Tags or Automate integration using our API.

Let’s get started! Feel free to contact us for a free demo.

Understanding eCommerce product recommendation engines in fashion

The time to ramp up product recommendations is now in fashion, and why AI is at the core of personalised shopping.

Creating a personalised experience for shoppers is a sure way of generating repeat business, building brand loyalty, making better business decisions through product analysis and as a whole, increase business longevity, producing a healthier bottom line.

Melissa loves social media.  Her preferred platforms are Instagram and Pinterest.  She has dozens of Pinterest boards and hops in and out of the app daily, looking for inspiration pinning multiple styles that reflect her personality.  Instagram, for Mel is all about tracking the most current hashtags and following her favourite influencers and celebrities.

90% of Pinterest users say using the platform helps them make a decision to purchase something.

What Pinterest does really well is provide highly relevant content based on user data and search queries.  Continuously learning about a user’s profile and preferences, serving up only the most inspirational content.

Going down the ’More like this’ path in Pinterest is like watching the makeover scene in Pretty Woman.

Reading any current ‘Future of Retail’ report and there are typically three key concepts driving change across the industry – convenience, automation and experience.  These studies are supported by large consumer groups surveyed about their shopping habits. One very common response is that shoppers are absolutely more likely to shop with brands that provide them a genuinely personalised experience, 71% actually, in the recent PSFK Future of Retail report and Shopper Attitudes 2019 Survey.

These statistics are not surprising. And, as everyday people, we can also appreciate what it feels like to receive a personalised shopping experience.  However, the pressure is on those working in the industry to create personalised experiences in their own business. This is where an understanding of empathy, nurture and delight are critical and should be prioritised just the same as resources and budget allocated to acquiring customers.  Why? Because the result of being empathetic, nurturing and delighting customers will organically convert into brand loyalty and repeat shoppers.

Back to Mel for a moment.  She’s aware of the latest trends in fashion, doesn’t purchase unnecessarily and will take some time researching and making a decision to buy something new based on what she currently owns, loves to wear and matches her style.   Mel has interacted with various fashion sites recommending products while she’s browsing the site, however she doesn’t want to see what others bought with the dress she’s looking at. She’s different. Melissa expects, at a very basic level that her individual fashion preferences are taken into consideration when being recommended products to purchase.

Technology has become significantly smarter in the past 10 years with artificial intelligence now the bedrock of many programs.  A number of platforms having already integrated AI to their software without users realising, namely, Pinterest and Instagram.  This is largely due to an exorbitant amount of data generated online.  From clicks, bounces, double-taps, actions and purchases to capturing browser history, downloads, uploads, streaming and geo-tracking.  Big data, our digital footprint, the way we behave online is feeding AI systems all over the world. It’s then translated into insights from patterns and analytics, which can then be used to complement business intelligence and improve the user experience.

When it comes to shopping online, most eCommerce stores will have a product recommendation engine installed.  This software carries out one main task, to recommend additional product to a shopper as they’re browsing a website and product page.  There are three types of product reco engines:

  1. Collaborative filtering – tracks and groups other users actions to predict what another shopper will like.  
    1. Eg: Others also purchased these products with this item.
  2. Content specific – a profile is created by tracking website pages and clicks of that user and the profile is compared to a product catalogue to identify which products to recommend.
    1. Eg: Products similar to this item.
  3. Hybrid – a mix of both collaborative filtering and content specific.

AI driven product reco engines are more sophisticated in that algorithms continuously learn over time users personal preferences and show products based only on similar tag and attribute characteristics.  This form of recommendation outperforms generic engines because it is highly personalised and programmed to adapt to preferences. This is significantly more useful than the logic of comparing a shopper to a group or catalogue in that one online shopping instance.  

One afternoon, Mel is enjoying a coffee and taking a few moments to scroll through Pinterest and casts her eye across an advertisement.  The ad image is a dress similar to a few images she’s pinned to a recently created Pinterest board labeled ‘Special night out’. Mel clicks the ad and is taken directly to the dress listed on a fashion eCommerce site she didn’t know about.  She reads the dress description, and will keep that top of mind while she continues perusing the site. Mel won’t make a purchase just yet, but she really loves this site and starts spending more time on it looking at shoes, accessories and back to dresses.  Over the duration of a few weeks, Mel is ready to buy the dress and clicks back to the site. She notices the recommended products are actually items she likes, and hadn’t noticed the earrings suggested previously. It was without hesitation that Mel could envisage the whole outfit together.  The dress with those gold statement earrings and she would wear her gold strappy heels that she’s had for several years and have never failed her.

The power of a personalised product recommendation through AI in Mel’s experience isn’t the technology, it’s her imagination bringing the outfit to life.  Mel is different, and so is everyone else looking to be inspired to shop and regular product recommendation engines just won’t cut it.

Fashion retailers of all sizes can now access these engines economically and as a standalone software product that isn’t part of a convoluted all-in-one package.  The team at Okkular can look after the implementation, getting you setup with their AI product recommendation engine and get your business on the path to personalisation.