By making product discovery easier for its users, the company lifted the homepage CTR by 193.5%.
The company in this case study stopped being a traditional telecom operator a long time ago. It is successfully transforming into an ecosystem that already spans more than 30 services — many of which prospective customers may know nothing about.
One of the company’s priorities is the active online promotion of its ecosystem products, and that makes it critical to offer users the most relevant, personalised options. To achieve this, the marketing team turned to Gravity Field.
A short path on the homepage
It was important for the team to introduce the company’s ecosystem products to the users who land on the website homepage.
Personalising the homepage is one of the most common tasks we solve for our clients. Gravity Field algorithms are good at picking the most relevant offers for users based on their behaviour in previous sessions. This case, however, was anything but typical. Most of the traffic on the company’s homepage comes from unauthenticated users who arrive from organic search, and we have no data on them yet. So we had to show first-time visitors the services they would find most interesting without knowing anything about their interests to begin with.
As a solution, the Gravity Field team proposed launching a campaign with dynamic allocation of the service icons on the homepage.
Dynamic allocation is the Gravity Field engine built on machine learning algorithms (known in the data science community as “Multi-Armed Bandit” and “Contextual Bandit”). It automatically serves the most relevant offer to every user at any given moment. At the very moment of interaction, the engine picks the offer or interface variant that is optimal for that user: first, by taking their personal interests and targeting rules into account; second, by running the test automatically and selecting the variants that optimise the communication in real time against a chosen metric — online conversion, revenue, CTR and so on.
This approach combines A/B testing with personalisation in a single run.
In real time, the Gravity Field engine selected the most useful services for each user and ranked them by the likelihood of conversion — products that previously might never have caught the eye became personalised and far more visible to every individual visitor.
To gauge user interest in the services, the brand team and Gravity Field looked at the widget CTR. Once dynamic allocation went live for new users — those the platform had no data on yet — it became clear that visitors were moving through to product pages more often. The numbers back it up: CTR uplift reached 94.9%, and for returning users 193.5%*.
*Results compared with the control group, for which the icon order was fixed.
Smart cross-sell for smart devices
Another ambitious task the brand team set for Gravity Field was to build personalised cross-channel communication, so that users would get a convenient, unified and seamless experience across every touchpoint with the brand’s website and its sections.
The company runs several websites built by different teams. Beyond the regional subdomains, whose job is to inform users about the full product range, there are numerous storefronts and product subdomains — an online store, for example. For Gravity Field, it was important to design effective communication for visitors across all of those domains.
Together with the client’s team, Gravity Field designed a cross-sell mechanic to promote the company’s mobile plans on its website. Among the most visited pages on the site are those for children’s smartwatches. The client team and Gravity Field anticipated what users would need next and placed a recommendation for a smart-device plan on those pages.
And the other way round: on the page for that plan, visitors saw the most popular children’s smartwatches. Users who activated the plan could buy those watches at a discount.
Thanks to the improved user interface, buying a relevant product became easier — add-to-cart for children’s smartwatches grew by 3.58%. Paired with the watches, users also took a keener interest in the plan itself: the CTR of the “I want this plan” button rose by 6.71%.
This city needs a hero
Encouraged by the results of the previous campaigns, the company decided to personalise its boldest, most attention-grabbing widget format — the hero banners on the homepage, the part of the site users notice most. That handed Gravity Field an even harder task: present digital products to subscribers and non-subscribers alike, keep traditional products just as easy to explore, improve the customer journey and grow the share of users.
Having analysed user behaviour in the previous campaigns, we concluded that it differs radically by region. So for the hero banners we decided to use predictive targeting, based on Gravity Field machine learning algorithms. Here, predictive targeting let us split users into segments made up of several regions sharing common behavioural patterns, and personalise the campaign settings for each segment.
Predictive targeting is a machine learning engine that analyses the results of every A/B test run on the platform across all possible segments, and identifies the user segments that respond differently to the variants in the test — whether those are promo campaign mechanics, visual elements, UI/UX hypotheses, creative, recommendation algorithms, merchandising rules and so on. Then, instead of settling on the single variant that wins “on average” in the A/B test, the engine automatically finds the optimal “test variant — user segment” combinations and targets its optimal content or communication at every user.
The Gravity Field engine split all of the company’s regions of operation in Russia into 6 core segments, and the team launched a slider of 4 banners with dynamic allocation for those segments.
As a result, users in Omsk, Perm and St Petersburg saw a cashback offer in the first screen of the homepage at the moment it was most relevant to their segment; users in Moscow, Krasnodar and Mari El saw an offer for Music and KION; and visitors from Kaliningrad, Karelia and Komi saw a banner with unlimited internet, TV and mobile service as a gift.
The campaign proved most useful for residents of Chukotka, Yaroslavl and Yamal — banner CTR there rose by 74.2% compared with the control group, which saw the default banner order.
Results
The company’s campaigns proved that the best way to promote ecosystem products is a personalised approach powered by artificial intelligence and machine learning. And even on a limited amount of data, it is possible to pick the most relevant offers for every user at any given moment. The telecom team managed to make product discovery more convenient through personalisation, improve the user interface and the customer journey, and put user segmentation and high-precision targeting to work.
In turn, the approach based on predictive targeting algorithms significantly maximised the performance of the campaigns that were launched. The engine surfaced distinct segments whose behaviour differed fundamentally from “average” user behaviour on the site, and helped build a separate customer journey for those segments, optimised specifically for them across the entire website.
The gains in numbers
+193.5%*
CTR uplift delivered by dynamic allocation of the service icons on the homepage
*Results compared with the control group, for which the icon order was fixed.
+6.71%
CTR uplift on the “I want this plan” button, delivered by the cross-sell with children’s smartwatches on one of the company’s plan pages
+74.2%
CTR uplift delivered by dynamic allocation of the hero banners on the homepage
We have become an innovative digital company and a visionary of data-driven communication with our customers. Internally, we adopt the very latest technologies and services to make the customer journey as effective as possible. Within the strategy we have chosen, that means the best customer experience through a deeply personalised approach built on artificial intelligence, big data analytics and a single customer ID. Gravity Field solutions help us deliver on that strategy. Together with the Gravity Field platform we were able to grow the share of users of our digital products and get people interested in our company’s ecosystem.
