Reading time: 5 minutes
Tags: e-mail marketing, internet marketing, online shopping
What the article is about:
- We will tell you how targeted marketing and machine learning help the growth of LTV (Lifetime Value) – an important marketing indicator that demonstrates the amount of profit per customer for the entire time of interaction with him;
- Why customer personalization is impossible without segmentation;
- Why revenue and email subscriber base growth shouldn’t always be your only target metrics.
Who is this article useful to?
- Email marketers, marketing directors;
- E-commerce marketers;
- Large, medium and small business owners;
- A business that wants to set up personal communications with customers and extend LTV.
Starting point and setting targets
Mario Berluchi is a Russian manufacturer of shoes, bags and accessories. In numbers, Mario Berluchi is:
- 20 years on the women’s footwear market;
- 4 offline stores;
- 200 thousand unique visitors per month in the online store.
In 2019, the brand faced stagnation, and it seemed that the revenue ceiling had already been reached. The team set a goal for themselves: to grow LTV metrics through personalization. For this, it was decided to combine all the available data on the company’s customers for end-to-end communication across all channels into a single database. It’s hard for a marketer to do this without an automated tool. Thanks to the CDP (Customer Data Platform) on the Mindbox platform, the Mario Berluchi brand managed to combine data from different sources (CRM, website, mailing lists) and build a complete customer database.
With the help of CDP, the department collected and unified all data from the purchase history, loyalty programs, and interactions with mailings. In the profile of each client, the entire history of his actions was visible: what he bought, when, from what source, the size and fullness of shoes, how he reacts to discounts, from which device he makes purchases.
Ivan Borovikov, founder of the marketing automation platform Mindbox:
“Marketing personalization is impossible without automation and aggregation of customer data into a single profile. This requires thoughtful goal setting: before implementing personalization technologies, you need to determine the metric that you plan to influence. In this sense, Mario Berluchi’s approach is close to me: colleagues have chosen a measurable metric – LTV growth – and are systematically working to increase it. ”
After 3 months of work in the customer data platform, the following results were obtained:
- Revenue from the email channel increased by 21%;
- LTV grew by 18.7%;
- ROMI from the email channel was 5636.6%.
Below we will talk in more detail about the 3 components of the work, due to which we managed to achieve such results.
Success factor # 1 – segmentation
The first success factor was personalization through user segmentation in email newsletters. It’s not just the person’s name in the subject line or header. Each fact about a client is an opportunity to build personal communication and increase loyalty. Mario Berluchi tried to collect a large amount of data and use it in communication.
For example: In an email newsletter with a selection of products or seasonal recommendations, a brand only sent the right size shoes to customers. However, size is not the only criterion. The brand also had additional information: foot width, fullness, city of residence. Information was collected from purchase history, surveys, tests and quizzes on the site. This scenario allowed us to triple the conversion to orders, its value amounted to 2.98%. On average in the market, the numbers vary from 0 to 1, which proves the effectiveness of the method.
Success factor # 2 – metrics
Most businesses take the standard approach of seeing revenue growth and subscriber growth as their top targets. At some point, the business hits the revenue ceiling, the subscriber base grows and begins to “burn out”. The marketing team realized that to find new growth points, you need to look deeper. Like most online stores, the company collected email addresses from forms on the site.
If a user has not made a purchase within 30 days, he will automatically fall into the “Old subscribers” segment. This is the largest segment of potential buyers who are not yet ready to make a purchase.
The flow of users from the “Old Subscribers” segment to “Customers” became a key metric to work with. Several variants of trigger chains were created for the “Old Subscribers” segment, which further allowed to increase sales.
Azamat Tibilov, Marketing Director of Mario Berluchi:
“The issue of correctly chosen metrics deserves special attention. Let’s say you consider the growth of your email subscriber base as a target metric for your business. You launched a promotion, your revenue went up, but along with it your unsubscribe rate went up. Is it correct? Difficult question. What happens to new subscribers? How do they behave during new sales? That’s what’s really interesting. ”
Success factor # 3 – machine learning
When a client visits the site, he leaves a lot of data about himself. The company decided to record all the actions of visitors in order to build a prediction algorithm using machine learning. The algorithm predicted the probability of a user’s purchase, the time after which the purchase will take place, whether the customer will add the product to the cart, whether he will return to the site, if he does, then when. Based on this data, the brand built personal communication with customers.
Ivan Borovikov, founder of the marketing automation platform Mindbox:
“Centralizing customer data, segmentation based on RFM analysis, personalized selections based on customer interests, predicting customer behavior using machine learning algorithms are signs of a mature approach to personalization.”
The principle of the algorithm is as follows: if the probability of a purchase is high, it is necessary to bring this user to the purchase without offering a discount. If the probability of a purchase is low, a pop-up trigger was triggered on the site. Let’s give an example: if, according to the prediction algorithm, the probability of a purchase is below 30%, then a pop-up with a 10% discount was triggered on the site. The user is afraid of losing such an opportunity and the likelihood of a purchase increases.
For customers with a high likelihood of buying who received a triggered email chain, the results are as follows:
- Open Rate – 45.91%
- Click Rate – 16.7%
- CR in orders – 4.35%
Issuing pop-ups on the site also gave good results:
- ARPU increased by 36.5%
- The share of abandoned carts decreased by 17.2%
- Purchase conversion increased by 16.5%
Conclusions and recommendations
- Personalization is impossible without segmentation, and automatic segmentation without CDP;
- A strategy and a well-defined plan is needed;
- The right metrics are the growth point for marketing;
- Analyzing data with machine learning will take your business to the next level.
He is fond of dancing and photography.