About the client
feeCOMPASS is a tech startup established in 2018. It develops customer service tools that combine the knowledge of sales specialists and the potential of artificial intelligence. It supports e-commerce sites by offering them products such as:
- needs survey – a short survey which is able to come up with individual product recommendations (usually 5 to 6 products) thanks to familiarity with the existing product range and by learning about the client’s needs and preferences,
- AI chatbot – a tool that makes it possible for the client and the brand to communicate in natural language (i.e. the language we use on an everyday basis); the chatbot is able to choose an appropriate product and explain why the recommendation is right for the client,
- recommendation engine for business – a service that provides shop owners with information about the lack of the product that would fulfill the needs of a certain group of customers.
At present, feeCOMPASS operates in the Polish market and its solution can be used by businesses with different profiles – from e-commerce selling electronics through beauty stores ending with online wine sellers.
The main challenge
The majority of e-commerce still operates based on a model that came to being around 2000. Websites feature lots of products, parameters, descriptions, filters. Customers are simply unable to get the hang of it all. With 100 products to browse through, they usually quit, make random choices, or… spend two weeks to learn and become experts in the domain of washing machines or TV sets.
Piotr Buszka and Artur Marsy started feeCOMPASS in response to the difficulties faced by online stores and their clients. The main problem is the unfamiliarity with the product portfolio and the difficulty in finding the right product, which translates into insufficient sales levels.
The frustration is usually most common among customers who want to buy:
- a product that needs to fulfill their individual needs – for instance, a cosmetic for a particular skin type,
- a product that comes in many different variants and with a hardly understandable specification – usually electronics and home appliances.
The popularity of e-commerce sites grew rapidly in 2020. Unfortunately, almost half of new customers isn’t satisfied with the shopping experience because the choice is too big and it’s hard to find the right products. These are the problems which feeCOMPASS has addressed with its needs survey, chatbot, and recommendations for business.
feeCOMPASS’ founders, having years of experience in managing and developing businesses (and well familiar with cloud-based solutions), knew exactly what challenges a tech startup would have to face. In the times when the popularity of online selling and shopping was growing (and the startup’s AI product was becoming more and more advanced), the founders had to consider the following factors in particular:
- limited resources – a problem experienced frequently by startups, one that involves a search for optimal solutions, a minimization of redundant activity, and requires process automation,
- a need to develop quickly – because the market is becoming more and more demanding, and even a short ‘idle time’ may result in being outdistanced by the competition,
- making sure that the product is stable and available – one of the most crucial points of agreements (SLAs) concluded between feeCOMPASS and its clients is the high level of the service provided,
- high scalability – feeCOMPASS constantly attracts new clients – companies of different size and with different expectations; the product is developed with a view to entering foreign markets, which is why a startup needs to be fully scalable – in terms of both business and technology.
GCP services enable us to scale the system in a stress-free way, which translates also into business scalability. Thanks to the high availability of the infrastructure and to the monitoring of service performance, we have the confidence to engage and work with big clients – who have big expectations.
The solution to the problem
Piotr Buszka, in charge of technology at feeCOMPASS, in order to meet market demands and provide clients with the best quality possible, chose Google Cloud Platform. He focused in particular on the key areas covered by the solution:
- Google Kubernetes Engine makes it possible to easily manage clusters involved in handling internal processes or test and development environments, a separate cluster with a production environment, and create new clusters and environments by means of a single command; the ease of creation of new environments translates into fast product development;
- infrastructure stability combined with undemanding operation requirements – at feeCOMPASS, Piotr is the only person in charge of the infrastructure; he admits that he spends very little time on its maintenance – the solution is very easy to use and most processes can be automated;
- app monitoring – the possibility to monitor the performance and availability of apps and services in the production environment as well as in other environments, notifications, and uptime measurements enable feeCOMPASS to provide its clients with SLA at the level of 99.99%;
- scalability – in the GCP environment, scalability is something natural; the app can be scaled horizontally by adding new instances or changing the configuration of a cluster; GCP supports also business scalability, when it comes to e.g. entering foreign markets – the app may be highly available to users from any part of the world;
- authorization and authentication services – feeCOMPASS’ employees use also G Suite, and Piotr can easily verify the identity of users and manage access.
I have lots of experience with cloud solutions and comparing Google Cloud Platform with other options, I’m really happy with what it offers. GCP focuses much on solutions like Kubernetes, which provides the necessary support in areas such as high availability or rolling updates. This is a great value for a startup.
Google Cloud Platform supports feeCOMPASS’ development in the following areas:
- development speed thanks to the availability of Google Kubernetes Engine,
- high availability and the possibility to monitor the performance of apps and services,
- focus on growth instead of on infrastructure maintenance,
- assurance of scalability of technology and business scalability support,
- assurance of high quality, availability, and stability of the solution, regardless of the size of the site and the requirements of the client feeCOMPASS gets to work with.
The value provided by feeCOMPASS to its clients – collaboration with Creamy.pl
The convenience that comes with using GCP translates also into the quality offered by feeCOMPASS to its clients. One of the sites that have become highly successful thanks to teaming up with the startup is Creamy.pl – a store selling natural cosmetics. Creamy.pl offers dozens of cosmetics for different skin types, ages, imperfections, or the preferred care time or expected results. Many new clients found it difficult to navigate through the rich product portfolio – that’s why the owners of Creamy.pl decided to make using the site easier by adding a needs survey and featuring (quite recently) an AI chatbot.
The survey consists of 7 questions concerning e.g. the skin type, the problems with skin, or the purpose of the product searched for. As a result, the client gets 6 recommendations of products or product sets suited to their individual needs.
Meanwhile, the AI chatbot lets customers communicate in natural language, helps them choose the best cosmetics for them, and explains the recommendations provided. The virtual assistant interacts with customers on the website level and via Creamy.pl’s Messenger.
The best proof of the success of Creamy.pl and feeCOMPASS’ partnership is the significant increase in the conversion rate among the visitors of the store. The conversion rate for new customers, who have used the needs survey, is 35% to 85% higher than the conversion rate for those who haven’t used the survey (the value fluctuates depending on the marketing activities undertaken).
We could see the results already two weeks after we started working together. In September 2020 we incorporated a new solution into Creamy.pl’s Facebook fanpage – an AI chatbot. The chatbot, thanks to its excellent knowledge of the product range, of each product separately, and of customer preferences, is able to ‘hand-pick’ the right cosmetic and explain why a given suggested product will be the best choice. Both the chatbot and the survey are as effective in selecting the right products as a real product consultant would be.