— is an Italian manufacturer of underwear and beachwear. They have over 100 thousand customers and an online shop established. Thanks to product recommendations, INCANTO increased its total revenue by 5.5% relative to the control group, and the email channel began bringing up to 19% of the online store’s total revenue, with an ROI of […]
INCANTO: an impressive 519% ROI from automating marketing, and that’s not counting additional revenue from product recommendations
30 Oct. ‘19— is an Italian manufacturer of underwear and beachwear. They have over 100 thousand customers and an online shop established.
Thanks to product recommendations, INCANTO increased its total revenue by 5.5% relative to the control group, and the email channel began bringing up to 19% of the online store’s total revenue, with an ROI of 519% provided by Mindbox. We will share statistically significant results of some AB-tests and feedback on the work done by the director of e-commerce at INCANTO, Andrey Lebedinsky.
Key figures
up to 19%
of the total revenue of the online store
is brought in monthly
by email newsletters
5.5% of revenue
earned monthly by INCANTO
INCANTOoff of on-site product recommendations
Relative to the control group
519%
ROI from Mindbox
Results
Two indicators are used to measure the results of working with Mindbox.
- The first is the revenue share of email newsletters relative to the total revenue of the online store.
- The second is additional revenue from on-site product recommendations relative to the control group.
Based on the profitability indicators of the email channel and the revenue from recommendations, we calculated the ROI obtained by Incanto from Mindbox.
Revenue from product recommendations on the website relative to the control group
To measure the effectiveness of product recommendations on the site, we set up a control group. Half of the visitors to the INCANTO website fell into the main group and saw product recommendations, the second half was put in the control group and did not see product recommendations.
We measured the difference between the two groups after 2 months. According to the results, visitors who saw product recommendations brought in 5.5% more revenue with a statistically significant probability of 95%. Unfortunately, INCANTO weren’t able to disclose absolute numbers. In the screenshot below, you can see that the value of a control group session is $0.95-1.05, while for the main group, which saw product recommendations it was $1.05 – $1.10.

Share of revenue from newsletters relative to the online store’s total revenue
The monthly share of newsletter revenue relative to the total revenue of the online store grew on average over the year from 12% to 15%. Data from Google Analytics, based on the last click measurement method. The graph turned out to be a bit floating. During the wintertime, sales were falling, and by summer they started increasing.
Revenue of the email channel relative the total revenue of the online store INCANTO

ROI from Mindbox
We calculated the ROI received from integrating Mindbox. In order to calculate, we took yearly indicators between June 2018 and June 2019:
- Margin of orders from the email channel based on the last paid channel attribute.
- Mindbox subscription price, excluding VAT.
The client requested that we do not disclose exact figures in this publication. The formula is as follows:
ROI = Margin of orders from email channel / Mindbox subscription price, excluding VAT × 100%
INCANTO’s ROI from Mindbox equates to 519%. Every $1 invested in the Mindbox platform brought the client $5.19 of additional revenue, and that’s without taking into account the funds from product recommendations on the site.
Director of e-commerce at INCANTO
INCANTO opinion
«Mindbox is an excellent tool that allows you to quickly start optimal communications with customers. All kinds of newsletters, product recommendations, and personalization can all be obtained through a single service and with proven efficiency. Products are being finalized according to given requests, the integrations are simple and clear, we have received no complaints regarding the interfaces from our email marketer.
Separately, I must note the excellent customer support, it was always prompt and fulfilled all of our queries.
Mindbox opens up so many opportunities for effective communications with our customers, so much so that we likely don’t even use half. Nonetheless, we will accelerate further and increase the number of newsletters, recommendations and personalizations in the store’s turnover».
The situation
Prior to migrating over onto Mindbox, INCANTO used a different platform for sending manual newsletters and single automated campaigns. There were no product recommendations on the site letters. The Mindbox platform was needed to:
- Increase the share of the email channel in the scope of the online store’s total revenue.
- Install product recommendations on the site and increase the rate of conversions to an order from product cads.
- Connect new communication channels for single-window mode operations.
Tasks
Based on the situation at hand, we identified several tasks:
- Run product recommendations on the site and test them with the control group.
- Automate communications via the email channel.
- Improve the performance of automated campaigns with the help of AB tests.
We will go over each task in further detail for you.
Launching product recommendations on the site
Hypothesis
Customers will make purchases more often if they see useful recommendations on product cards. On the INCANTO site, product collections are configured according to the personal recommendation algorithm: we recommend products which are viewed and purchased by customers that are similar in their behavioral characteristics to one another.

Website option versions
To measure the revenue from on-site product recommendations, we set up an experiment through Google Optimize: 50% of the site’s visitors belonged to the main group, and were exposed to product recommendations on the site and in the product card, the other 50% visitors landed into the control group, and did not see product recommendations.
Test results
Our hypothesis was confirmed. The differences are statistically significant, with a confidence level of 95%. We saw a big monetary difference around 3 months after the start of our experiment. It turned out that product recommendations on the product card brought in 5.5% additional revenue for the online store relative to the control. In the screenshot below, it can be seen that the value of the control group session is $0.95-1.05, while in the main group, which saw on site product recommendations it comes out to $1.05 – $1.10.

How communication via the email-channel were automated
In total, 11 automated campaigns were launched in a year. Among them were:
- Subscriptions for out of stock goods.
- Abandoned cart.
- Next purchase recommendations.
- NPS survey.
Let’s go over two most popularly converted communication workflows: subscriptions for out of stock products and abandoned cart.
Out of stock subscription
If a customer is interested in a product that is currently out of stock, he can subscribe to it and receive a notification when the product is refilled. The chain consists of two letters. The first letter arrives if the product has not returned to stock in a week. The second letter arrives when the product is back on sale.
The first letter goes out with a recommendations for similar products, if the origin product is not yet available for sale.
If the product was not listed available in a week’s time, we will send a selection of popular products that are available on the site, based off of product similar product characteristics that a client has previously enjoyed.
The second letter goes out when a becomes available
The second letter is transactional, it arrives regardless of subscription status. If notified the customer that a product which he is subscribed to returns back for sale We recently launched this mech, so no recent stats are available.
Newsletter | Open rate | Click rate | Conversion to order | Average order value |
---|---|---|---|---|
Recommendations of similar popular products | 61.7% | 18.5% | 0.203% | $76 |
Abandoned shopping cart
The abandoned shopping cart chain is launched if a customer has added goods to his cart but did not complete a purchase. The chain consists of three letters, the first arrives within thirty minutes. Following this, letters are sent with a promotional code and a reminder about said code, but only if the client has not executed an order within a week.
Within the abandoned shopping cart chain, we suggest recommended products that are similar to ones contained within the cart.
Newsletter | Open rate | Click rate | Conversion to order | Average check |
---|---|---|---|---|
First letter regarding an abandoned shopping cart | 39.6% | 12.9% | 0.357% | $90 |
Second letter with a promotional code | 13% | 12.9% | 0.499% | $87 |
The third letter contains a reminder about the promo code | 47.1% | 14.7% | 0.528% | $93 |
To increase conversion statistics for automated campaigns, we set up and launched some AB tests. Let’s go over two AB tests that we carried out.
Test # 1: «Bestsellers on sale»
Hypothesis
If a «Go to site» button is added within the automated letter covering bestsellers at a discount, then clients will transition to the site more frequently. Thus, the CTR (click-through rate) increases.
Newsletter options
There were two options: one with a «Go to site» button and without a that.

Test results
The hypothesis was confirmed. The differences are statistically significant, with a confidence level of 95%.
Letter option | % clicks |
---|---|
With the «Go to site» button | 4.4% |
No button | 3.7% |
The difference in clicks was 0.7 %. Thus, substituting an additional button for transitioning to the site increases the CTR by ~20%.
Test # 2: Abandoned shopping cart campaign call to action
Hypothesis definition
Clients will open emails more often if they see a call to action in the subject line.
Newsletter Options
There were two options:
- Letter subject, «Not ready to buy?»
- Letter subject «Not Ready to Buy? Better hurry!»

Test results
The hypothesis was confirmed. The call-to-action letter opening rates are higher with a statistically significant difference.
Letter variant | % openings |
---|---|
«Not ready for a purchase yet?» | 38.05% |
«Not ready for a purchase yet? Better hurry up!» | 39.82% |
The difference in discoveries was 1.77 %. Thus, a 5% greater call to action rate increases open rates.
Conclusion
Over the last year, we launched 11 automated mechanics and conducted a series of AB-tests to improve performance. The monthly email channel share grew by an average of 3% according to Google Analytics. We installed a product recommendations widget on the website, which increased monthly revenue by 700 thousand rubles relative to the control group.
The next step is to launch web push newsletters and integrate Mindbox with the INCANTO mobile application when it is launched.