Birchbox: using data for personalized discovery

Birchbox uses data science to create value for customers and brands


The company

Founded in 2010, Birchbox is a discovery commerce platform that offers consumers a personalized way to discover, sample, shop, and learn about the best beauty products and brands out there. Katia Beauchamp and Hayley Barna (HBS graduates) founded the company as a subscription service that would curate makeup boxes with sample assortments each month. The company has since added a full e-commerce platform where customers can buy full sizes of their products they sample, and opened its first brick and mortar store in New York City.

To date, Birchbox has raised over $70M in funding, operates in six countries, has nearly 300 employees, over 800 brand partners, and more than one million subscribers. Annual revenues are estimated to be around $100M.

Value creation

Birchbox creates tremendous value for both sides of its platform: brands and customers. On the brands side, sampling has been used as a marketing tactic for decades, but with lackluster results. Generally speaking, non-directed sampling results in samples that are the wrong color or for the wrong skin tone/type. In addition, brands are unable to evaluate sampling campaigns because they don’t have data to do so. Birchbox fixes this problem by sending samples to customers with the correct profile (e.g. if you are a teen, you won’t receive anti-aging cream). Birchbox is able to extract ratings data from its platform to understand user preferences, and even see if the user proceeded to buy the full-size item from the e-commerce store. The subscription service creates a new channel for these brands, and funnels new customers their way.

On the customer side, each subscriber fills out a detailed profile before receiving her first box (e.g., hair type/texture, skin tone, style preferences). Through Birchbox, customers have the opportunity to try high end beauty products that are tailored to their individual features, personality, and style, without having to commit to the expensive, full-sized version.

Value creation using data analytics

Birchbox has a robust data science team that drives value creation. For brand partners, Birchbox provides valuable insights, data, analytics, and consumer research.

For customers, Birchbox’s data science team sees each monthly box delivery as an optimization problem with the following parameters: profile preferences, limited supply of samples Birchbox can send, policy to never send the same sample twice, etc.

The company also understands the high cost of sending samples that the customer does not like, and has processes in place to involve the customer as much as possible. For example, Birchbox gives customers the choice to choose one item per month they can include in the box. This serves as one of Birchbox’s highest engagement points with the customer, and allows customers to know what to expect.

Birchbox also uses data for personalized discovery. Through its online platform and mobile app, Birchbox will make personalized recommendations based on customer ratings and order history, using key words like “discover,” “based on your profile,” and “the latest.” At the same time, Birchbox’s data science team takes care of the customer through the many touchpoints the company has: web, offline, brick and mortar, mobile app, and social media. The data science team has developed specific tracking systems to follow customers’ interactions with the company and show the right messages at the right time, thereby enhancing engagement and conversion.

More recently, Birchbox has launched its own makeup brand called LOC. The products in this collection focus on bright colors that are on-trend and seasonal. Birchbox undoubtedly used its vast warehouse of customer data in product development.


Value capture

Birchbox has three major revenue streams:

  • Subscription revenue: $10/month for women and $20/month for men per box (represents ~75% of sales)
  • E-commerce market: users can buy full-size products directly on the Birchbox website without a subscription
  • Affiliate model: Birchbox charges brands some fee for introducing their products

Looking ahead

Going forward, Birchbox will continue to meet growing copycat competition in the subscription beauty product space. Smaller competitors such as Glossybox, Goodebox, and Ipsy’s Boxes have emerged, while larger players like Sephora have also been entering the market using a similar business model. The Birchbox founders believe they have a competitive advantage over smaller competitors because Birchbox has its full e-commerce store while competitors do not. The e-commerce store creates more value for brands as it helps deliver on Birchbox’s promise to deliver new, long-term customers. An e-commerce store is much harder to develop because it requires inventory, customer service, and more complex logistics relative to the simple subscription business. At the same time, Birchbox has also been acquiring smaller competition (JolieBox) to gain scale and entry into new markets.



Pymetrics: Play Games, Get a Job


Netflix: Bringing Data Analytics to a Creative Industry

Student comments on Birchbox: using data for personalized discovery

  1. Does Birchbox’s launch of its own private brand create a conflict of interest on their platform? What has been the response from other brands who might get displaced by LOC? It seems like a gutsy move when you’re playing the role of an aggregator. Sure they probably have great data on what people like and took advantage of it (you might compare this with Netflix creating original content), but in a business more dependent on brand loyalty for repeat purchases, it might be a bit too aggressive.

  2. We recently had a BirchBox case in Entrepreneurial Finance, and I must say I’ve been thinking a lot about the company since the case. After reading your post, it is clear to me how consumers benefit from data, but I’d be curious to hear more about what type of data the company’s request from BirchBox. I assume they want to know at the very baseline what people are buying, but do they only get data as it relates to their brand? Do they receive site trend data overall? What data do they find most valuable? Do they get this type of data from other retail partners?

  3. I initially had the same concern as JG – that introducing their own line of makeup might create a conflict of interest, but I’m actually not sure that’s true. Sephora sells its own line in its stores alongside other brands with great success. Netflix producing its own content is another great example. Instead, the two biggest concerns I have are 1) what is the purpose in opening a brick and mortar store? Does the store collect information on who comes into the store, what they buy, etc., or can people pass in and out anonymously? Unless they manage to sign up / create a profile for everyone who comes through the store, I wonder how they can capture useful data here. 2) I agree that there is some value creation for makeup brands in effective sampling of their product to drive full-size purchases, but I am curious to know what the conversion rate looks like. It seems like some may subscribe to Birchbox for a monthly surprise gift box and may not actually be interested in purchasing full-size items.

  4. Thank you for your post. I think one of the interesting things that the company does in terms of big data is the match between the detailed profile a customer sets up at the beginning of the subscription and the personalization of each box. The question is the accuracy of the data provided, remember that the output of a model is only as good as the inputs. If a woman does not understand her skin type, or hair needs the products received can be the wrong ones. An interesting example of a beauty company trying to improve the inputs is Sephora. They provide in person consultations supported by technology such as color skin recognition to understand the best make up colors for your skin. Information that can be later used for targeted marketing. I wonder if the brick and mortar store is a step to improve this experience. In addition, I wonder if they can create and capture more value from brands, for example giving them more feedback and working together to come up with products that solve customer pain points.

  5. Recently, I have been thinking a lot about the role of technology in creating meaningful customer experiences in retail and Birchbox is a great example. Although this is a beauty discovery service and it is important to have a wide range of samples, to a large extent, the real value behind Birchbox is in its algorithms and data capabilities and the products are commodities.

Leave a comment