Task
To develop a service, which will select a personal vitamin set based on person’s health condition. Service is oriented on women from 25 to 40 y.o. Users made their way to the app by Instagram advertising or from blogger’s links.
Project info
Project team – 4 people
Budget – 13 500 $
Timeline – 6 months
Technologies – Django, React.js
Features
Wellbe is a start-up. Which means that we made many iterations of the product during development process. We made changes in the project iteratively – once a week. During the working process, we changed its concept 3 times and made a redesign once.
Solution
We launched project on a market, where were two similar products. As a basis of CJM and sales schemes we took experience of competitors. We were developing, testing and evolving our product.After half a year of development, we came up with following userflow:
1. User is taken to landing
2. Fills in a
questionnaire, answering questions about their lifestyle
3. Service generates a personal recommendation consisted of 3-5 vitamins
4. User sees price of
vitamin set, and, if needed, removes extra vitamins
5. User specifies delivery address and then pays for an order
Service in pieces
Questionnaire
The goal is to create a questionnaire, that will cover all life areas. There were more than 100 questions. Not all of them were shown to user. For some answers a clarifying question was asked. Question format was different: multiple choice, manual input, picture choice. It was necessary to develop our own questionnaire constructor, that freely allows to adjust content of questionnaire, move from one question to another and calculate results.
Service in pieces
Questionnaire
The goal is to create a questionnaire, that will cover all life areas. There were more than 100 questions. Not all of them were shown to user. For some answers a clarifying question was asked. Question format was different: multiple choice, manual input, picture choice. It was necessary to develop our own questionnaire constructor, that freely allows to adjust content of questionnaire, move from one question to another and calculate results.
Why do we need to create our personal questionnaire constructor?
Together with client’s specialists we created our own unique algorithm of vitamin selection. There was no ability to connect to it already existed constructors.
Algorithm of vitamin selection
Based on our constructor, doctors created a questionnaire consisted of 140 questions. In order to turn user’s answers to recommendation we developed a three-stage algorithm.
1. It was designed in such way as not to harm the client in any way. At the slightest danger of drug mismatch, it removes from set.
2. Every element of algorithm can be customized. Value of answers, correspondence matrix and special recommendations can be changed fast via admin panel.
3. Not only algorithm allows to select vitamins, but also to test marketing hypothesis. Depending on a situation to the selected set may be added related products.
Personal recommendations
After processing of user’s answer, algorithm generates a personal recommendation. Our goal was to convince user that the recommendation is trustworthy and that they certainly can help to solve user’s problem or can help to reach user’s goal. And of course to convert user to buy set. It was important to show that service do not make recommendations, but select and sell personal vitamin sets.In order to achieve these goals, we:
Designed a recommendation page in landing page style, in order to make it more official.
Implemented “why?”, that explain why certain vitamin was selected to personal vitamin set.
Placed vitamin set price on a separate page to move user for purchase.
Where does the e-commerce start?
After launching failure of first version of a product, where user was moved from recommendations right away to ordering page, we understood that we need interim page. The purpose of such page would be to meet users, who is interested in purchasing. Something similar to cart at online shops. The page of vitamin set “adjustment”, where user would be able to regulate vitamins in set, to know a final price and to enter a promo code.
For payments we implemented Cloudpayments widget as it not only supports payments via bank cards, but also Apple/Google Play pay method.
After order was completed, user makes a way to telegram-bot.
Inside features
Analytics
We strived to track every step of our users. We used Yandex. Materica, pixel and investigated user action logs. In the end we have a big excel table, in which we can see next data: from where did user find the app, what user answered in questionnaire, what orders placed, whether user used promo code or not and etc.
Pivot table was formed in auto mode. To get such report, we only needed to choose report time interval in admin panel.
Marketing tools
The data, which was collected from questionnaires, was further used for remarketing.
We did an integration with SendPulse. After the questionnaire, a chain of mails and SMS massages was activated. The chain was customized by marketing specialist in a separate interface.
Results
We successfully worked with our client during full life cycle of start-up. After 6 months of working, our client was able to attract 100 000 $ investments. However, after some time project was closed. Now, we and our client continue working together on new projects.
