If you haven’t heard of it, Thread is an online-only fashion technology brand with a unique proposition.
Instead of sending you generic product suggestions or ‘trends’ emails, each week a personal stylist will deliver product recommendations based on your preferred styles, color palette, body type, budget – and your feedback on previous items.
But here’s the twist: despite servicing around 650,000 customers (all men – its offering for women is in the works, but not yet live), Thread employ fewer than 10 stylists.
So how does such a small brand also manage to personalize its product to each and every customer?
Personalization at scale
I know what you’re thinking. And you’re right. There’s some fancy tech sitting behind this.
Thread stylists use a machine learning algorithm called Thimble to help make recommendations that are tailored to you as an individual. Stylists and Thimble work in tandem to deliver hyper-personalized emails to users, explains Tom Banham, Content Editor at Thread:
“Thimble helps our stylists do what they do at scale. The machine learns your tastes…and it can either help to broaden your horizons, or it can help you find things you know you like, so you don’t have to sift through thousands of product options to find them.”
“It enables [our stylists] to maintain the client base they have, but still make the service feel personal and useful.”
The level of personalization they offer is certainly compelling. The signup process presents me a series of selection tasks designed to assess my preferences.
I click on brand logos I typically buy from, choose styles I like, and answer a couple of lifestyle questions like ‘do you wear a suit to work?’ and ‘which outfit would you wear to the pub?’ (Can you tell they’re based in the UK?)
The welcome email I receive tells me how Thread works and informs me I have been assigned a personal stylist, Kasia…
…before she follows up with a personal introduction:
So far, so personal.
“The way we see it, the more personalization we can offer the better,” says Tom, “Our aim is to be a platform for everyone, not to be purely for people who don’t like shopping, but also for people who are super fashion-engaged. The challenge is to service all those people in a way that feels specific, and feels personal.”
I ask Tom what role email plays in this journey.
“Email is certainly a big part of it. We obviously have our paid acquisition through social media and through search, but our emails are designed more for our existing users. I personally think that when you’re in someone’s inbox, that’s a responsibility.”
“It’s a very personal thing, and you shouldn’t just be sending stuff because you feel like you need to be present. We want to make sure all of the content we send out genuinely improves the lives of our users.”
Tom explains that Thread sends out three types of email:
1. Stylist recommendations
These are a core part of the service – a link to a curated list of items alongside a personal message from your Stylist.
“There needs to be a balance here. We don’t want to be too prescriptive, because we understand people have fluctuating tastes, they might change their mood…so we want to be able to offer people inspiration as well as pragmatic solutions.
2. Browse drivers
This email is designed to drive traffic to the ‘browse’ feature on thread.com.
“We try and make sure this is filtered around the weather, and not things like trends. It’s more practical.”
Sent out on Saturdays, this email contains links to relevant blog content, offering practical recommendations and advice that is designed to engage customers beyond the purchase.
“We want to offer customers more than just a place to buy clothes, but a place to improve their personal style and build a relationship with the things they wear.”
All Thread’s emails have small touches that make them stand out – whether it’s an eye-catching gif or a countdown timer for a sale, each email feels like it has a purpose.
What does the future look like for Thread?
While the content of Thread’s emails is already quite sophisticated, Tom revealed there’s more work to be done:
“Our content still isn’t as personal as we’d like it to be. We have the data, but not the resource to make it happen.
“We do what we can. We test subject lines, we test formatting to optimize for CTR and open rates. But the ideal for us would be to shoot seven different versions of every single campaign that we do, with a model of the age, style, body types and price bands of our users.
“But that’s the future. For now, we’re segmenting lists by category. Or we might send emails based on where a customer lives, or the weather outside. So at least it feels like you’re getting something for you.”
Of course, he says, it’s also important to avoid Shiny Object Syndrome, and ensure any changes you make are tied to improving ROI:
“There’s an infinity of options, so it’s also about trying to work out which are going to make a difference. The technology is so powerful it can be tempting to just do stuff because the option is there…and then you realize you’ve wasted two days building something that isn’t going to create any uplift in engagement.”
Where will email be in 5 or 10 years?
“I’m not sure. It’s had a resurgence in recent years – I think in part because it’s gone from being very generic to feeling very personal (if you can do it properly). Other channels like social have become so swamped with advertising and marketing, as a user you can feel like you’re just being blasted at.
“Whereas a channel like email – if you can get through the noise – you can create something that someone feels engaged with. What we’re trying to do is to achieve brand recognition in the inbox, so that when someone sees an email from Thread, they’ll want to open it no matter what the content is. The difficulty is, when you’re sending the more generic content you can get lost in the noise.”
“I think as long as brands can work out ways of creating relationships through email, it will continue to be super important. It will become more personalized and more specific to what people need at certain times. The fact is, people still use email in enormous volumes, so it’s always going to be somewhere you can contact people.
“We’ve seen really good engagement with email in a way that we haven’t always on other channels. So I think it will work in tandem with other channels, but it will play a slightly different role going forward.”
With so many fashion retailers looking to differentiate themselves, one can’t help but wonder whether Thread’s data-driven offering signals a change in the way retailers are approaching fashion.
As access to machine learning capabilities becomes easier for forward-thinking brands, will traditional retailers be able to keep up with the change?