With More Data Science Behind It.
Here’s where we explain how we get our data, and how we know it will never let you down.
Nothing Matters More than Accuracy
For data to be useful, it has to be right. It’s just that black and white. Our data’s one purpose is to work for you. We’re not lenient on ourselves, and don’t expect anyone else to be either. When you use EDITED data we want you to know that you’re looking at the real market, no blemishes or oversights.
A few members of our data team, the people committed to keeping EDITED the most accurate source of retail data in the world.
Working with so many smart people on something so challenging is great, but the real reward is hearing how much EDITED has helped our clients improve.
Sophie Coy, Retail Data Manager at EDITED
EDITED doesn’t just set the bar high for retail data, it sets the bar high of the entire data industry. I love being a part of that.
Paulo Sampaio, Data Scientist at EDITED
We rally behind the idea that data can make retail more efficient. Nothing makes us feel better than when we hear we’ve helped that happen.
Alejandro Giacometti, Data Scientist at EDITED
I take the trust placed in us by so many great companies very personally. When our customers use our data to do better, it’s incredibly satisfying.
Louise Brooke, Retail Data Specialist at EDITED
Taking something which has always been considered only for ‘numbers people’ and making it accessible to anyone is what inspires me to make EDITED the best.
Joe Berry, Data Engineer at EDITED
When you’re making commercial decisions based on data, trusting that data is everything. That’s why we make everything transparent. EDITED lets you drill down and see the numbers behind the numbers. You can trace every single data point back to its source and see for yourself. We also have the biggest data QA team in our space, so you can always rest assured that what you’re looking at has been checked and double checked.
Our data has been tested against the internal data of some of the world’s biggest retailers and has been proven the best in the business. Spot on in every category. Once you have access to EDITED, we welcome you to check it against your internal data and see.
Using Data is No Longer a Choice
Every company looks at data. But if that data is internal, or coming from an in-house system, it’s only half the story. To see the entire market it takes full teams of highly qualified people working non-stop on collection, analysis and storage. It also takes staying on top of new methodologies and advancements in machine learning. All things we focus on at EDITED so you can keep your focus on what you’re best at: retailing.
From day one, you’ll have the real-time information you need to stay ahead of the market and make huge financial gains from it. The product pays for itself. And it puts you on equal footing with the world’s biggest brands and retailers, some of whom have been using this sort of data for years. You’ll also be able to prevent mistakes that result in big losses. But we don’t like to harp on the negatives. It’s not about preventing the bad, but enabling the good.
Hear about how Farfetch use EDITED’s data every day (2:07)
Getting started with EDITED is easy. We’ve got a Customer Success team here to help you get the most out of our product at all times (email, phone, Twitter, up to you). You can be up and running the same day you become a customer. It means less work and more profits. It means the peace of mind that comes with knowing no one will ever know anything before you do.
How we collect and analyze our data
You’re relying on our data, so we think it’s important to be open and explain where it comes from, and how we make it work for you.
We crawl the web, exactly like Google and Bing do.
In the same way that Google uses machine learning to build crawlers that can read websites and understand all the publicly available information from them, we use ours to read the sites of brands and retailers all over the world. But reading is only half of it, the second — and more crucial — part is understanding.
Using advanced machine learning, we’ve taught our systems to do more than access and collect information, we’ve taught them to understand what they’re looking at.
In this way, our systems can do things like determine when a skirt is not a dress. Or when a tunic is not a shirt. It does this with near mathematical perfection for every product on the market, every day.
One way we do this is by image recognition.
That means recognizing a piece of clothing within an image and separating it from non-essential elements, i.e. the model wearing the clothes. Literally, chopping off arms, legs and heads to single out the item you want to see. As a means of double-checking itself our systems also read the item’s name, description and meta description to make sure the words correspond to what is in the image.
And it’s not as easy as looking for the word “dress” and categorizing it as a dress either. If it were, men’s dress shirts would be an interesting category!
It takes a massive computer vision and natural language ConvNet neural network models to make sure we remain the most accurate in the industry, and the world. If you’re not up to date on what a neural network is (who is?), let’s spend some time making it clear.
So what do we mean by ‘data point’?
As it corresponds to a real-life thing, what we call a ‘data point’ is just a product as it exists at a certain point in time. Let’s take a dress for example. Or more specifically, a dress at 7:01 p.m. on August 7. Our system captures that dress online and converts it into a data point. Contained within the data point are all the details about that dress, things like color options, price, drop dates and so on.
Every single detail about a product contributes to the formation of one complete data point.
Our system then finds the same product the next day and assembles its details to form another data point. If the first and second data point are not identical, the system knows to alert us to what has changed. Maybe it is no longer available in size medium, for example. We’d record that event in the product’s timeline. If two days later the data point shows that the medium is back in stock, we’d record that too.
We see and record every slight change to a product over time, so you know everything about it.
These data points are also rended visually. So in our dress example, we’d simply see the dress with all the relevant information alongside, instead of a pile of numbers. And because the dress’ data point is collected daily, we can actually track the evolution of the product over time.
For a single dress, it may take capturing hundreds of details to form a complete data point. EDITED compiles each of these details every day, and shows how each has changed over time. Then it does that for 40 million other products, comprising more than 330 million SKUs at 90,000 brands worldwide.
In this way we actually break down the entire global apparel retail industry every day and send it streaming back to our offices for processing. It’s by reconstructing these data points, totaling nearly seven million per day, that we’re able to begin building a real-time view of your market. Regardless of segment or country.
Say you were looking to compare prices of every red dress in the USA between $20 and $30, EDITED would do more than look for words matching ‘red dress’ - it would sort through the data points to find every relevant product, whether it was called red dress or not. In this way you can be assured you’re working with 100% visibility, backed by the freshest data.
As for details about in-store merchandise? We don’t need them.
Retailers have gone omnichannel. What’s in stores is also what’s online because it has to be. Customers won’t tolerate browsing in one place, and not being able to buy in the other. This self-balancing act means that what EDITED shows you is the real market anywhere in the world. And if we don’t, we’re always happy to work with our customers to add crawlers specific to their needs.
How we Contextualize our Data
Q: Why does context matter?
A: Without context, EDITED would look like a series of seemingly random numbers.
What brings our billions of data points to life is contextualization, the process by which our data team augment and restructure the numbers and put them into a real-world context. Or simply: turning the data points into pictures, charts, graphs, etc…
Everything you do in EDITED, be it creating pricing structures or analyzing replenishment rates, is made possible by contextualization. It’s the numbers transformed into the visualizations and insights you’ll use to guide your commercial decisions.
And it takes entire teams, with more than a couple specializations to do it and some pretty state-of-the-art equipment to do it right. Especially when dealing with such an incredible quantity of data. More than anything, it’s this expertise and ability to augment the data that sets EDITED apart from being just another collection of web crawlers.
We’re structured so that there are just as many people dedicated to contextualization as there are data collection. That means the attention we give this process is completely unmatched by anyone and results in a product that allows you to get the most out of the data, quicker and easier.
Who handles all this data?
We already mentioned our data team, the people who collect all this data and bring it to life on screen, but that’s not where our involvement in the data ends.
We’ve got an in-house group of analysts with decades of combined industry experience to watch the market for you. Our analysts are EDITED super users that prepare reports on everything from trends to category performance market-wide, all backed up with EDITED data.
We’re passionate about fashion and data. Our reports don’t just list must-have trends, they show you how to make them work commercially.
Emily Bezzant, Head Analyst at EDITED