ThirdLove looks to data science for support in meeting its inclusive retail goals
- Summary:
- The online underwear retailer is on a mission to provide a bra for every body shape, size and skin tone - and it’s looking to a cloud-based data warehouse from Snowflake to provide the insights it needs.
With its focus on comfort and inclusivity, ThirdLove is an excellent example of a new breed of bra retailer - one that’s working hard to understand its customers and what they really want. And, as the precipitous decline in fortunes at rival Victoria’s Secret suggests, what they don’t want so much anymore is push-up bras, modelled by super-skinny, largely white models, portraying a fantasy of overt sexuality seemingly aimed at straight white men.
By contrast, ThirdLove, founded in 2013, keeps a keen eye on modern realities. The online retailer was an early pioneer of half-cup sizes, in recognition of the fact that many women struggle to find a comfortable fit among more standard sizes. Its bras come in 80 different sizes, ranging from 30AA to 48I, compared to an industry-average range size of fewer than 40.
In terms of inclusivity, as well as embracing different body sizes and shapes, ThirdLove’s ‘nude’ shades come in a range of colors to match different skin tones, rather than the standard pinky beige that can only be considered ‘nude’ for white women. On the company’s e-commerce site, products are showcased by a diverse range of models.
In a 2018 full-page advert in the New York Times, CEO Heidi Zak summed up ThirdLove’s approach like this:
We believe the future is building a brand for every woman, regardless of her shape, size, age, ethnicity, gender identity or sexual orientation. This shouldn’t be seen as groundbreaking, it should be the norm. To all women everywhere, we see you, and we hear you. Your reality is enough.
Planning ahead
It’s a powerful message and it seems to be hitting the right note. Last year, the San Francisco-based company received a $55 million funding round from a group of investors that included Anne Wojcicki, co-founder and CEO of 23andMe; her sister and YouTube CEO Susan Wojcicki; broadcaster Katie Couric; and Nancy Peretsman, MD of investment bank Allen & Company.
At the time, ThirdLove said it would be using the money to build out its size range and extend its product lines. Data science will play a big role in understanding how best to do this, says Rebecca Traverzo, who joined ThirdLove in December 2019 as its VP of Marketing. A big believer in the power of data-driven marketing, she also heads up the company’s data science and analytics team. As she explains:
It’s really critical for us to understand the customer, her entire customer journey, how she behaves on the site, her interactions with our CX [customer experience], the products she’s buying, the reviews that she’s posting. To date, we have almost 600 million data points captured from over 11 million women to really create the best fitting, most comfortable bra and to adapt her user experience with customization and personalization opportunities.
A great deal of that data comes from the online Fit Finder tool on ThirdLove’s site. This is essentially a quiz that uses machine learning to help women identify the correct band size, cup size and style for them, based on what other brands they wear, in what sizes, how they fit and the common problems they experience, such as slipping straps, gaping cups or a band that rides up in the back.
As its platform for this kind of analysis, ThirdLove has been using Snowflake for about a year now, she says. Prior to that, it was using the Amazon Web Services (AWS) Redshift platform:
But as we grew, and found we had more and more data points and much more customer knowledge, we really needed a more powerful platform, offering better tools and better performance.
Marketing tactics
Eight months into the job, Traverzo’s aim now is to build an approach that closely combines both ‘art and science’, she says:
So the data science and analytics team are the science side of this: they’re the information lever that we pull to drive all our other marketing functions, and that’s an integral part of my approach. But that being said, there’s also the art of all this, you know? So the marketing teams come in and take the insight that data provides but add to it the human element - a little bit of gut instinct and a lot of experience - to really understand consumer behaviour and what’s happening in the wider marketplace from an overall trend perspective. To me, it’s about rounding out data-driven insights with more human and compelling implications for our future marketing campaigns.
Podcast advertising, for example, has been a very effective route to market for ThirdLove. (Listening to Gretchen Rubin’s Happier podcast, for example, was the first time I personally heard about ThirdLove.) Says Traverzo:
It’s super-efficient - and that’s something we like. We’re very much tied to maintaining a very efficient CPA [cost per acquisition] in order to drive margins, especially in today’s COVID-19 world. We’re a successful start-up, sure - but we’re still a start-up, right? So we need to preserve cash and extend our runway, even with all the uncertainty with everything that’s happening in the world today.
So with podcasts, every show has a unique code that customers use so that we can track them from a redemption perspective. We’ve built our own proprietary marketing mix modelling [MMM] tool, so that we can ingest all of these data points into Snowflake through various API [application programming interface] integrations, along with data on Facebook marketing performance, Google performance and so on. That means we can use this MMM model to identify where we would put our first marketing dollar - which channels perform best for us, especially from an incremental perspective.”
This is important, she explains, because if she and her team can identify the incremental gain in sales that might be obtained by increasing spend on a particular podcast or social media site, they can better optimize their marketing budget. Or, to put it more simply:
We’re not just throwing more money into a channel or platform that would basically drive itself.
For the future, Traverzo is interested in integrating more data from the Snowflake Data Marketplace, where Snowflake customers can acquire third-party datasets to combine with their own, internal data. In ThirdLove’s case, that might be datasets relating to fashion apparel in general, as well as category-specific datasets relating to trends and patterns in bra sales:
Being able to gain more insight into what’s happening not just in our own business but in the wider industry would be incredibly valuable for us. It would enable us to do some more benchmarking at our end to make sure that the trends we’re seeing internally are good or need improvement, when they’re compared to the wider industry.
That's important because while ThirdLove is already doing much to redefine an industry that is newly realigning itself with the needs of ‘real’ customers, like every company in this space, it will need to say agile, while still keeping that strict eye on margins.
According to fashion industry analysts, comfort and functionality are now firmly in the spotlight, with sales of padded and push-up bras falling, and sports bras and ‘bralette’ styles rising in popularity. New options in skin tones, sizes and sustainable production are all increasingly important in terms of winning a share of customer budgets. Younger consumers are certainly less responsive to the traditionally sexualized imagery of the industry. Gender-fluid lingerie options are starting to emerge.
There’s still a long way to go - and success in this very different marketplace will require careful planning for every company involved. ThirdLove, it seems, already has its foundation in place.