Introduction
Hi, I’m Kelly O’Brien from Findify and this is our Tech Talk.
So, Findify! Findify offers a powerful artificial intelligence solution which puts the products shoppers are most likely to buy – right in front of them – when they search, browse collections or act on recommendations. This delivers a personalized product experience – which dramatically increases sales for ecommerce merchants.
While Findify has its origins in Stockholm, Sweden, the company has hubs across 9 countries worldwide. Personally, I’m working out of Ireland right now, while still working closely with team members in Germany, Sweden, the UK, and Russia.
I’ve been working with Findify for just over a year now as their Marketing Manager.
On a daily basis you could find me writing blog posts about Findify’s latest technological developments, showcasing our latest clients and interesting product customizations, compiling ebooks and white papers, or managing Findify’s active social media presence, all to help new and existing clients get the most out of the platform.
I am also, of course, in charge of creating educational and informational videos – like this one!
Introduce the market problem that your company/ technology solves
And the first thing I’m going to take you through today, is the main problem that Findify technology solves – Findability. We are solving the frustration of shoppers not seeing the products they are interested in – both when they search and when they browse collections.
Numerous studies have shown shoppers who conduct site searches are three times more likely to convert. There are several reasons for this but, most important is that these shoppers are focused on a desired item or category, and if they are shown a relevant match then the mental process of continuing the sales journey remains straight forward and can continue unimpeded.
In this situation the biggest threat to conversion is an interrupted flow – which can easily happen if the site search is poor.
By not showing the shopper relevant products when they explicitly expressed what they are interested in is a huge waste. It’s essentially leaving money on the table. You spend a lot of time, money, and effort attracting people to your store with marketing efforts, only to lose them at the final hurdle with an underdeveloped site search.
Market Problem
Without a doubt, the biggest challenge facing ecommerce merchants is the battle for attention. Shoppers have an increasingly short attention span and if you can’t capture them with relevant products, the risk is huge that they will go somewhere else.
If ecommerce merchants don’t show their customers relevant products on the very first page of results, studies show 75% percent of shoppers will go somewhere else.
Change
Delving deeper into those studies, we see that shoppers also have a huge bias towards the first products shown. 21% of all shoppers conducting a search will click on the first product shown, while 70% will click on one of the first 8 products.
What does this tell us? It tells us the first products you show have to be good. They have to be relevant. Getting even just this part right is in itself a great opportunity to boost sales.
What is your technology name, how have you positioned the company to solve the market problem.
So now you know WHAT you need to do – but HOW do you do it?
You need to make your site search smart – so that shoppers conducting searches receive relevant results. You do this by improving search capabilities by introducing advanced spelling tolerance, natural language processing, a zero results workaround, one-way and two-way synonym capabilities, things like that.
But you also need to go one step further and ensure your site search is catering for individual desires – showing shoppers not ONLY results relevant to their search term, but also relevant to them as individuals.
At Findify we do this by harnessing the power of artificial intelligence. Developers at the company have spent years creating, tweaking, and refining, our very own machine-learning algorithm which we have infused into our three main solutions: Personalized Search, Recommendations, and Smart Collections.
As soon as a shopper visits a website with this personalization, the AI algorithm starts learning about them. It learns, in real time, what their preferences are, based on their behaviour. The solution then reorders products based on these learned preferences, showing shoppers more of what they’re most interested in and, therefore, are more likely to buy. I always think of it as having an automated personal shopper for every single one individual shopper, which is actually pretty cool.
Personally, I’m very proud of the technology we’ve created here, but not only because it’s cool, but because there’s not a lot of other people doing this – and even those that are aren’t doing it with as advanced an algorithm as we are. So I get a real kick out of that.
Describe the history of the tech/company. What is the background of the company, how old, how many employees, journey, number of customers, valuation, type of ownership public/private.
To give a brief history of Findify, the company was set up in Stockholm, Sweden in 2014.
For the first five years of its lifetime, Findify was a company focused very heavily on product development – perfecting the algorithm and the solutions until they had a truly market leading product.
It’s only in the last year or so that Findify has truly started to market itself, spreading our name quite literally across the globe.
In the last six years, Findify has gone from a four-person start-up to a team of 20 highly skilled workers operating from 9 hubs worldwide.
To date, Findify solutions have been trusted by more than 1,800 ecommerce merchants all across the globe – with the majority of current clients being based in either North America, Europe, or Australia.
Who is best suited to use your tech, examples of current clients
While Findify solutions will improve literally any ecommerce store, the biggest uplift is seen when there are complexities within the store’s assortment.
So, for example, if a merchant only has 17 products on their store, advanced search and personalization isn’t going to be as useful as on a store with a larger assortment. Ideally it is from 200 products and upwards that merchants will really start to see the effects.
Some examples of current clients include UK fashion brand VictoriaBeckham.com, well-known sports apparel store Everlast, RocketDog footwear, BH Cosmetics, MJS Electrical, Reese Witherspoon’s fashion store Draper James, and Pharrell Williams’ Billionaire Boys Club.
Traditionally we’ve attracted a large number of clients in the fashion, jewellery, footwear, and cosmetics industries, however we do also create a lot of value for our clients in electrical, beverages, automotive, gifts, interiors, health and wellness, and many more.
How specifically do you solve the market problem.
As I mentioned earlier, our three solutions are personalized search, recommendations, and smart collections.
With Personalized Search, there are two aspects – search accuracy and personalization.
Accuracy is about making the search as smart as possible so that it can return relevant results to the customer putting in the query. This includes an autocomplete, predicting what they will type as they’re typing, meaning fewer keystrokes for the user.
Other aspects include Natural Language Processing – the search being able to tell the difference between products, like dress, shirt, shoes, and attributes, like the colour red or the size 14. An effective search should also allow you to search by SKU number, which is helpful for electronic stores where customers search for makes and models. Other aspects include spelling tolerance, ensuring customers get relevant results even if they make a typo, and a zero results workaround – ensuring relevant products come up even if the exact query searched for is not present.
Personalization is all about the AI algorithm which analyzes the behaviour of the customer and learns, in real time, what their preferences are. Then, when they carry out a search, they get relevant results thanks to search accuracy capabilities I just mentioned, but the results are now ranked in order of their preference, which was just learned by the AI algorithm. So if they have a preference for gold jewellery, for example, or for loud floral maxi dresses, this is what they will see more of.
Our second solution, Smart Collections, takes all the same aspects I just mentioned and applies them to collection pages to ensure products appearing within collections are ranked according to the customer’s preference, and are not simply a one-size-fits-all static list of products.
Our Recommendations solution also draws on those learnings, but to increase upsell and cross-sell capabilities using designated space to highlight products the shopper is most likely to buy.
What is your unique selling point (USP)?
Our unique selling point is, without a doubt, our personalization algorithm.
Our developers have quite literally spent years on this algorithm, tweaking it, perfecting it, making it the best it can be. And the thing is, it’s never quite done because every month, every year, there are still advances in this technology which we keep abreast of and infuse into the algorithm on an ongoing basis. Currently, the algorithm analyzes more than 220 factors that influence buying decisions.
We also regularly test our algorithm against the services offered by our competitors. What I can share around this is that our algorithm has proven superior to every single other solution we’ve tested it against, which is a huge source of pride within the company.
Further tests, which we can talk about, reveal the algorithm boosts ecommerce revenue by an average of between 7 and 29%. The increase in revenue depends on how good or bad the previous site search method was.
Demo of your tech
So what does Findify actually do? We have talked about the high level problems Findify is solving, but let’s deep dive into the product.
First of all, Findify personalizes the results for every shopper on the site. This is done by taking all dimensions related to each product into account. For example, category, brand, colour, descriptions, titles etc etc
These are then matched with the individual behaviour of the shopper. What products are they engaging with, where do they spend time, what do they put into cart and buy. By marrying their behaviour with all products dimensions, the algorithm can easily predict which products are most likely to be relevant for them.
This is done in real time and for every query. Which means that two shoppers with different preferences will be shown a totally different ranking for the same query, here exemplified with a shirt. And since they have a huge bias for what they are shown, this increased relevance will have a huge impact on their purchase decision.
Let me show you a simple example. Here I’m looking for a dress. By only giving the algorithm a few data points, interest displayed through clicks skipped products, I show a curiosity for visual products. Now, conducting a second search you can see that the algorithm in real time has started to rerank the products I’m shown. This is of course just a simplified example and these colours are not forever going to haunt me, but my profile will be more and more flushed out by every action, throughout the session and when returning.
Overview of key functions
Findify is powering the full product experience, throughout search, categories and recommendations. In the dashboard you’ll find actionable analytics and can curate the product experience with manual control. This is particularly useful for tactical reasons when you want to push a specific brand collaboration, push large stock levels or other objectives.
There are a lot of things I could show around search, but to not spend an hour I’ll keep it brief. Here you’ll see the autocomplete. Main purpose of the autocomplete is to give the shopper positive reinforcement that their query will lead to positive results, and hence lead them further into the funnel. You can also see our variant search capability, the same for autocomplete as results, where not only the most relevant product will be matched but also on a variant level to reassure the shopper they will find what they’re looking for.
Smart collections are bringing the ux and personalized ranking from search into category pages. I’m here showcasing the out of the box filter setup.
Meanwhile recommendations exist to drive cross-sells and upsells. This illustration might be a bit overwhelming but I usually explain it like a large lego box where everything is possible. Instead of starting with the individual parts, you should start with the business strategy. What do you need to achieve? For example: Higher margins, bigger basket sizes, closing shoppers with long sales cycles. By first establishing that, we can easily tailor the recommendations to fulfil the desire. And then there are endless opportunities to iterate and compare performance of different recommendations.
Examples of customers helped
Metrics are very important to Findify and we like to keep track of the impact we are making for our clients.
The most recent study we carried out was on our client Lisen.dk.
In this case, we enlisted the services of an independent AB testing company to really measure our solutions and see what value we were creating.
Listen.dk is a Danish fashion store, running on the Shopify platform and they have the full suite of Findify running on their store – personalized search, recommendations, and smart collections. We found that Findify solutions resulted in a 26 time return on investment for Lisen.dk.
There was a 101% increase in conversion rate, meaning an additional 5,000 products sold in the first year of implementing Findify. There was 8 times more interaction with recommended products, and an overall increase in revenue of 7.1%.
What was the before and after impact?
When clients start using Findify, they usually notice two main changes – their user experience is improved, and their conversion rate goes up.
In terms of UX, new clients usually remark upon how much better and sleeker their site’s search experience now looks. For many, the autocomplete has a lot to do with this. The autocomplete pops up as soon as a customer starts typing in the search bar and, depending on merchant preference, the autocomplete can have images, stickers, even add to cart options already at this initial stage. The merchants also notice their site search is more intuitive, and that the filtering options for search results and within collection pages are more robust.
In terms of an increase in revenue and in conversion rate, this is a hugely important metric for all of our clients, and one they watch closely. With Findify integrated into their site, a new merchant can see they are getting as much value as possible from each and every visitor to the site – using Findify site search, personalization, filtering, smart collections, and recommendations to ensure customers can find what they are actively looking for, as soon as possible, and also that browsing customers are given appropriate and relevant product discovery opportunities.
Any quirks or limitations of your tech?
At Findify we recognize that there are many ecommerce platforms used by merchants to set up their online stores. We have therefore focused heavily on making the integration process as smooth as possible. The technology is platform agnostic with a full front end library to use.
So far, our focus has been to master the product display on site. There are of course a lot of assets that can be personalized on site but we have decided to stay focused on products since that has shown to make the biggest impact for the merchants bottom line.
Where does your tech fit in the MarTech stack?
Findify is the shopper facing layer connecting your clients with ideal products.
Findify is platform agnostic, which means you can connect the service with any platform, custom setup or headless solution. Either through quick installed extensions, javascript or API.
We have built a complete front-end library that you can use out of the box, or customize to desired experience.
Findify’s role is to display relevant products for the shoppers, no matter if they search, explore collections or act on product recommendations. This is achieved through our own search wrapper – fully optimized for ecommerce, trend scoring – dynamically altering the products shown for new shoppers based on aggregated data, a merchandising suite enabling manual full control – and, most importantly, personalized products based on the shoppers behaviour on site.
Since Findify powers areas that display products across the site, it has been very important to build a solution that is flexible and can be fully adapted to the clients brand experience. This offers endless opportunities through customizations and pre-built integrations such as powering Yotpo’s reviews throughout the product experience.
Pricing, onboarding and timeline of adoption
In terms of pricing, we have three main plans. These are the Premium, Professional, and Enterprise plans. These can be accessed at www.findify.io/pricing.
Where a merchant fits in our pricing plan depends on which solutions they want to access, but also on usage – how many visitors their site attracts each month.
After selecting a plan, our team will talk to you about your requirements and mark down any customizations you’d like us to implement. While you can have Findify installed over a warm cup of coffee, any desired customizations will take between 2 and 4 weeks to create.
So there you have it! A full rundown of all things Findify – our company history, our clients, our best-in-class solutions, and our powerful AI algorithm. If you’d like any further information, you can go to www.findify.io or email [email protected].
I’m Kelly O’Brien from Findify and thanks for coming to our tech talk.