VamSvet: how to achieve an ROI of 1135%

Our first success story with a decorative lighting store. We’ll tell you how in a year and a half of operation the share of the email channel reached 7% of the company’s revenue. We’ll share the feedback from VamSvet marketing director Evgeny Sukhanov and Tatyana Kuznetsova, leading email marketer at Kokos Mailing. We also talk about testing product recommendations with unexpected results.

Results

To measure the results from working with Mindbox, VamSvet uses two indicators: the share of the email channel relative to the company’s total revenue, and ROI.

ROI from Mindbox

We calculated the ROI from Mindbox. For this, we took yearly indicators from May 2018 to May 2019:

• The Margin from orders from email channels based on the last click attribution.
• The Mindbox subscription price, excluding VAT.

The client asked us to not disclose exact figures in this publication, the calculation formula looks like this:

ROI = Margin from orders from the mail channel based on last click attribution / Subscription price for Mindbox, excluding VAT × 100%

VamSvets ROI from Mindbox is 1135%. Each ruble invested in Mindbox, brought the client almost 11 rubles!

Share of email channel relative to total company revenue

The average annual share of mailings relative to the total revenue of the company increased from 3.7% to 6-7% .Data was provided by the customer. Method of measuring used: last click in Google Analytics. A control group was not used: VamSvet marketers did not want to exclude any part of their customers from their communications.

Share of revenue from the email channel,%

The situation

VamSvet was working with a mailing service provider through which it was possible to send only basic triggers. There was no real-time transfer of orders available. Instead, they were downloaded once per day. The Next Best Offer mechanics did not work optimally, recent orders were not taken into account. For example, a client was recommended fixtures that were not suitable for the chandelier purchased by him earlier on the same day. In product recommendations, the only algorithm that worked was one for substituting popular products, for which it was impossible to define categories or assign priority.

Kokos Mailing created a development strategy for VamSvet, which implied more sophisticated mechanics, for example, triggered mailings while lowering the cost of an item in cart. Mindbox was chosen for their implementation, because the platform was able to realise all marketer needs.

1. To set up complex customer communication chains.For example, send out reminder emails to those who open a letters, but who did or did not go to the site, and bought or did not buy goods. Stop such reminders when an order is placed.
2. Set up AB-testing of communication efficiency.
3. Use custom product recommendations in emails.

How to automate communications through an email channel and test them using AB tests

The project has 49 automated campaigns covering the entire life cycle of a client, from a subscription confirmation to the welcome chain to stimulating contact for repeat purchases and outflow reactivation. They provide more than half of the income from the email channel – 63.6%.

Revenue share growth from triggered mailings

There were 9 groups of triggers launched on the project, including welcome chains, abandoned cart, abandoned view, reserved products, Next Best Offer. Let’s talk about some of these mechanics.

Abandoned cart – price decrease

In addition to the standard abandoned cart mechanics: a reminder of an incomplete purchase and a discount offer – a more complex trigger mailing is configured on the project. It is sent in case a goods price decreases by 10%.

Abandoned viewing of an expensive product

This newsletter is sent if a customer has viewed a product for more than a certain amount, but has taken no further action. Similar goods are substituted into the body of the letter, but are a little cheaper.

Reactivation chain for orders

Consists of 4 letters, used on customers only once.

• The first letter is a reminder from the company about new products.It’s sent to customers who made a purchase more than 1 year ago.
• The second letter contains a promotional code for 500 rubles and is sent three days after the first, that is if the client has not made an order in that time.
• The third letter is a reminder about the promotional code.It’s sent after three days, if the customer has not made a purchase.
• The fourth, a survey letter, is sent to customers without any purchases five days after the previous one.

A promo code is the best way to reactivate clients

How the effectiveness of automated campaigns was evaluated with the help of AB-tests

Tests were conducted in the abandoned cart and abandoned viewing mailings. Let’s go over one of them.

According to the rules, the AB test must be used to check a single element from the letter. In this case, VamSvet specialists decided to test the header, as it affects the open rate of an email.

The second header version, the emoji one, won. It’s average check was 1.2 times higher. This allowed the company’s marketers to evaluate which header increases the open and click rates, as well as leads to greater conversions. The differences are statistically significant ,with a confidence level of 95%.

How to use product recommendations in triggered mailings and test them using AB tests

After integrating with Mindbox, VamSvet gained the opportunity to test and apply various algorithms in letters.

Hypothesis definition

It was assumed that there was a greater likelihood that clients will make orders for goods that were most bought by others. From this we deduced that it would be best to substitute in the most purchased products instead of the most viewed ones. This mailing was sent to customers who did not open letters from VamSvet over the last 90 days. As a target action, we selected the transition to the site.

Test results

The hypothesis was not confirmed. According to our test results, it turned out that emails with most viewed products turned out to be more effective – this algorithm provided more links to the site, and the average check for this option was 1.8 times greater. The differences are statistically significant, with a confidence level of 95%. This proves that any, even the most logical of hypotheses, need to be tested well. The next step is to verify the effectiveness of personal recommendations.

Conclusion and what’s to be expected next

Since the start of the project over a year and a half ago, we set up 49 automatic campaigns, which bring in more than 60% of the email channels revenue. They cover the entire life cycle of a client and generate profits from each new subscriber. In the future, we will be conducting AB tests to better fine tune triggers.