case study: measure everything

18 May ‘16
logo is an online furniture store that customers like for its convenience and optimal pricing. They previously used 3 different email marketing systems, which was complicated and costly. In January 2016, they switched to Mindbox.

We transferred all mechanics in one and a half months, which cut costs by 60%. Over the following three months, we ran several more new campaigns, conducted an interesting recommendation test and optimised the process for sending manual emails.


  • Thanks for your order
  • Abandoned view
  • Abandoned category
  • Abandoned cart
  • Reactivation
  • Special campaign for regular customers
  • Suggestions for next purchase and several more recommendations

GA transaction rate grew from 1.3% to 6.3%


  • Review us
  • Welcome sequence
  • Activate doubters
  • Additional services

GA transaction rate grew from 1.1% to 10.3%

Optimised manual email creation

Every product used to be placed in emails by hand. Now, the right products can be placed directly from the product page filter into the template using any conditions. This reduced email preparation time by 78%.


Recommendation test

The abandoned view emails have a group of viewed products, as well as a «just for you» group. This group is comprised of products generated by a recommendation system using a special algorithm. has product ratings based on the number of views on their site. We wondered what would happen if we made our own simple recommendation algorithm and compared it with the professional algorithm?

We conducted an A/B test. We used identical email templates, texts and subject lines, but in 50% we used the group of products from the recommendation system, and in 50% we used products chosen randomly from the list of most popular products from the category viewed.

The test currently shows no difference. There is no statistically significant difference in GA transaction rate, average check or click rate.

Simple Professional
2% ±1,3% по GA 1,6% ±1,1% по GA
1,6% ±1,1% по GA 8900 р ± 6000 р
12,3% ±1,3% CR 11,8% ±1,1% CR

What could it mean?

  • Perhaps there’s a difference, but it’s less than the current margin of error. It would take about a year to tell.
  • Perhaps the recommendations don’t work in this particular campaign – only familiar viewed products draw attention.
  • This might be specific to the particular industry or store.

One thing is certain: you need to measure everything. What works in one situation can easily fail in another. The test was quite interesting, another is currently underway, and several new campaigns and tests are in the works.

Dmitry Cheryukin
Dmitry Cheryukin
Head of Marketing at

«Before working with Mindbox, we used three systems. Data were stored separately, and the systems’ potential was limited. In January, we transitioned smoothly to Mindbox. After the integration, we stopped using third-party trigger services and automated the emails that we previously sent out manually. We were surprised: our costs fell by 60%, while email revenue doubled thanks to the new automated campaigns. During the switch to Mindbox, we centralised our address database and boosted the number of customers that we could send emails to by 1.5 times! We aren’t stopping there: we plan to run and test new trigger-based campaigns. Summary: We’re happy with the service and highly recommend it! »

Mindbox team:

Filipp Volnov
Project Manager

Semyon Mikanyov
Lead Manager

Ilya Vasyutovich
Technical Manager

Tell us a little about yourself

We’ll respond within 24 hours

Partnership request

Typically we’re answering within 24 hours