Personalization has become a popular buzzword for marketers – more and more customers expect marketing messages to be directed specifically to their interests and needs. For marketers, getting personalization right is a process more than an end-goal. As we are able to gather more data about customers, the opportunities for personalization grow exponentially. Today, personalization means a whole lot more than customizing your emails to include a recipient’s first name!
We talked to Meghan Maupin about how Atolla is creating truly personalized skin care treatments for their customers using machine learning and data science.
Meghan Maupin | Interview
After completing a Bachelor of Science in Architecture, Meghan began her career doing 3D printing at an architecture firm, before spending two and a half years as a Senior Designer for 3D printing startup Formlabs. While there, she was inspired to become more knowledgeable about early-stage business and what it takes to bring a design to market.
Keen to merge her experience in design with an education in engineering and business, Meghan enrolled in MIT’s Integrated Design and Management (IDM) Master’s program. Her startup background inspired her to perhaps become an entrepreneur herself. While at MIT, she found the intense stress of schoolwork had a debilitating effect on her skin and the traditional avenues for skincare simply weren’t working.
This her gave her an idea…If only there were a scientific approach to skincare: a method that would enable users to select skincare products based on their personal skincare needs – what works, what doesn’t, what products might cause allergic reactions, and more.
So in August 2019, Meghan, along with her data scientist and dermatologist co-founders, launched custom skincare brand Atolla.
Can you tell us a little bit more about Atolla?
Atolla is aiming to reinvent the skincare industry. 1 out of 3 Americans have had an allergic reaction to their skincare products, yet people still don’t know what skincare products are right for them. To make matters even more difficult, when making this choice you also have to factor in each person’s unique genetic makeup, a host of environmental factors, and the fact that skin also changes over time — so it’s a extremely tough problem to solve.
Personally, I think that the entire skincare industry is broken right now — it currently focuses on marketing and branding rather than efficacy of a product on an individual level. Brands seem to have totally forgotten about the end-users themselves.
The Atolla team currently has 8 members (we’re hoping to reach 10/12 people towards the end of the year) and we’ve been doubling in size month-over-month since the beginning of 2020. Before we launched, we raised a pre-seed round of funding and we’re currently closing out our fundraising for the seed round — so it’s exciting to see new investors coming on board to help Atolla during this next phase of growth.
We’ve launched one product so far: a custom serum that helps people treat specific skin issues and adapts to their changing skin needs. We create the serum from data we collect about their diet, lifestyle, environment and their skin (including their oil, moisture, and pH levels). Every month, users do an at-home skin test, and we use the results to refine their serum formulation. In the next 6 months, we plan to introduce new active ingredients for wrinkles and fine lines, as well as custom moisturizers and cleansers.
How does your data model work?
Atolla has a patented process for skin analysis (combining objective and subjective data) and custom skincare product formulation. Each month, we refine and optimize a user’s formulation because our process is set up like a giant feedback loop. We have thousands of users testing their skin each month and helping us build (and calibrate) our models. The good thing is that the more data we have, the more effective we’ll be at predicting exactly what someone’s skin needs. We’ve been collecting skin data since we began building the technology over 2 years ago at MIT, so we’re finally at the point to see some significant trends!
Machine-learning and data science are at the heart of everything that we do; the data will always guide us to what products to launch next that meet our user’s skin needs.
What does your approach to personalisation look like?
There are 3 steps, built around a monthly feedback loop:
1. Start: Answer a few questions about your skin history, concerns, and lifestyle/environment. Our data-informed system will then create a custom starter serum to address your immediate needs.
2. Refine: Measure your skin at the end of the month using our Skin Health Kit. Input results directly into your profile — there’s nothing to send back.
3. Adapt: Our patented algorithm gets to work designing your next serum, using your physical skin data and any noted updates to your lifestyle or environment.
We not only provide people with a baseline for their skin health, but we also look at how this changes over time — that’s something that our users find particularly interesting.
What does this all look like from the end user’s perspective?
Our first interaction with our customers is to gain a better understanding of them, their environment, diet/food intake, oil, pH, and moisture measurements. We provide feedback during this initial process: letting you know if you are within the ‘healthy’ zones or if there’s some discrepancy that you should watch out for.
This data goes into creating your own custom formulation and we also explain what specific ingredients should work for you and how that relates to your skin measurements. We show users a health dashboard where you can see how you’re tracking against your goals and if you’re beginning to get into your ‘healthy’ zone. This dashboard also includes instructions on how your custom Atolla formula fits in your entire skincare routine, and allows you to keep track of how your other non-Atolla products are performing.
Our insights section provides even more information as you continue to test and use the products. However, it’s not all tech-led; we also have a human element. We have our own in-house esthetician who complements the information that we already have — she can go through your dashboard and all your data with you before explaining everything in more detail.
How are you doing this all with such a small team?
We actually founded the company all the way back in November 2017, but we spent two years building the technology and refining our process before we launched. Our technology is the key to our business — it’s our competitive advantage — so we needed to make sure we got it just right: training the model, training the data, and figuring how to make skin tests possible at home with your mobile phones.
We are lucky enough to have ridiculously smart and talented people on the team who are uniquely qualified to work on this. We follow an agile framework, which enables us to work on different aspects of the digital and physical products all at the same time. Lastly, we’ve made sure that we are vertically integrated, so we control manufacturing (which again helps us learn exactly what percentage of what ingredient in the product is working for your skin). This integrated, cross-functional set-up has been absolutely crucial to our success — it gives us full control over the entire end-to-end-process.
Signup for our newsletter, participate in future articles or get help with your Data, Analytics, Product Development, and Marketing Automation. Signup Here →