How Petrovich combined all of their customer data into single profiles and increased their direct marketing revenue by 2.5 times

8 Nov ‘19

Petrovich— is a DIY ecosystem, consisting of a website, an application for iOS and Android, 19 retail stores, a call center. Also, it provides a loyalty program for builders and a “Professionals exchange” a recruitment system, where professionals and clients can find each other. In 2018, the company’s revenue exceeded 49 billion rubles. Their back-office is built on 1C.

We created a single client profile for Petrovich. As a result, we decreased costs on the creation of different letter variants as well as integrated some interesting tools:

  • An elaborate scheme of mailings over 5 channels, launching different triggers depending on the client’s real-time actions.
  • Communications on the Digital-room project: clients create a design project offline, and send themselves an email with items from the online store.
  • Dynamic content in mailings for 11 different cities.
  • Product recommendations for repair work orders on the «Professional Exchange».

A single profile helped increase the efficiency of these such mechanics.

Key figures

2.5 time

increase in revenue from
direct marketing
(excluding the impact on
offline sales)

3 fold

coverage increase (number of
customers who opened at
least one newsletter per
month)

5 channels

used instead of one
for communicating with clients

774 thousand rubles

savings per month
by getting rid of an agency

Data received from the client via Google Analytics using the last-click attribution method

We will go through every figure below.

Audience Reach

After we personalized content, the number of subscribers that opened at least one newsletter per month began to grow. An example of such personalization on the project is the welcome-chain separation into 2 two options: for professional construction workers, and for clients buying goods for themselves.

Audience Reach

Direct marketing revenue

The key business indicator for Petrovich marketers is revenue growth from direct marketing. Based on plans for 2019, it was supposed to double. The results, however, exceeded all expectations: the channel showed that revenue increased by 2.5. Data from Google Analytics based on the methodology of the last paid channel:

Direct marketing revenue

Reduced the cost of creating mailings by stopping work with an agency

Using dynamic content for 11 cities, we were able to get rid of the agency that was previously used by Petrovich and hire an in-house marketer. Which turned out to be much cheaper:

Reduced the cost of creating mailings by stopping work with an agency

12 905 rubles of savings on each newsletter. That means 774 thousand rubles of savings for 60 newsletters per month.

Maria Bratchikova
Maria Bratchikova,
Direct-communications Manager at Petrovich

«Together with Mindbox, Petrovich began to develop direct marketing significantly. The central aspect of Petrovich as a company is the service and convenience provided for clients, and we were able to transfer these qualities onto our customers through newsletters.

We successfully merged all essential info in one place and set up communications across various channels. Our messages are unobtrusive and relevant to the interests of our customers. From mass mailings, we moved on to in-depth segmentation, we take into account a vast number of contact points, and work on both offline and online behavior. We are building communications with all clients, regardless of where they came to us from.

I would advise Mindbox to those who are already collecting customer data and need automation, segmentation, and personalization».

What we have been done to achieve results

  1. Created a single customer profile.
  2. Made communication for customers simple: two sites as one.
  3. Developed an omnichannel scheme of commercial mailings.
  4. Personalized messages based on a client’s city.
  5. Automated and personalized mass mailings: AB-tests, control groups, and Next Best Offer on machine learning.
  6. Made the customer omnichannel experience simple.

After integration, Mindbox became a master system: it consolidates data from various sources so that clients interact with all Petrovich services on different channels and do not feel a transition between them.

Created a single customer profile

Source of data for the single client profile
Source of data for the single client profile

Mindbox combined data on clients that were previously stored in several places: by Petrovich themselves and contractors. A single customer profile contains the entire purchase history: online, offline, as well as call center and mobile application purchases. This allows us to find out how many unique clients the company has, place restrictions on the frequency of newsletters, so as to not spam clients, but to receive reliable information regarding the efficiency of communications.

ОOne client identifies himself in various contact points, including offline
One client identifies himself in various contact points, including offline

Information on interaction with all of the company’s services are sent to Mindbox: page and category views on the site, orders on the «Professionals exchange» and orders through the call center. Each client is unique; this is due to automated data deduplication.

We made communication seamless for clients: two sites as one

A single client profile makes communicating with Petrovich seamless. After a client leaves a request on the Petrovich «Professionals Exchange» for a master to perform construction work, he receives a confirmation letter. Depending on the type of construction work chosen by the client, Mindbox automatically recommends the most popular products for this specific type of construction work. There’s a separate set of the most popular goods for each type of construction work. The letter opening percentage is ~ 50%.

A letter with product recommendations motivates customers to make purchases on the main Petrovich site
A letter with product recommendations motivates customers to make purchases on the main Petrovich site

There are other emails in the chain, including a reminder of unread responses from masters, which only a third of recipients open it.

The letter arrives if the client has not visited the site within a day of the masters response.
The letter arrives if the client has not visited the site within a day of the masters response.

We developed an omnichannel commercial newsletter scheme

Communications with clients and personalisation will become more complex as the company aggregates more data.
Communications with clients and personalisation will become more complex as the company aggregates more data.

Petrovich stores all customer information in the Mindbox CDP. The Mindbox consulting team was able to develop an omnichannel communication scheme. Tasks included:

  • Not spending a lot of money on expensive channels while retaining the same coverage.
  • Communicate with each client through a convenient for him or her channel.
  • Not repeat info to a client on all channels without a good reason.
  • Take into account actions at all points of contact (online, offline) in each communication channel.

Such a scheme is impossible without combining online and offline actions in one place. Data must be enriched in real-time to rebuild communications while taking into account the behavior of clients across different channels – for example, viewing a product on the website or buying an item in the retail store.

The customer base is divided into three «branches»:

  1. Clients that installed the mobile application.
  2. Clients interact with Petrovich only through email.
  3. Clients with nothing but a telephone number tied to their profiles.

If the client has a fully filled profile, communication begins with mobile pushes: the cheapest and fastest communication channel. The next priority channel is email, followed by web pushes, and finally Viber. If we know only an email address, then we send letters, if it’s just a phone number, then we send messages through Viber. If nothing is known about a client, except for his consent to receive web pushes, then we will interact with him through the website.

Thus, Petrovich can communicate with their entire client database, selecting the most convenient communication channel for each client and not excluding those who have only a mobile application installed or have provided just a phone number.

Personalized messages based on a client’s city

Petrovich has retail shops in 11 cities across Russia
Petrovich has retail shops in 11 cities across Russia.

There are Petrovich Construction centers in 11 Russian. Prices for the same products differ in some regions. This depends on suppliers and logistics costs. Inventories are not replenished simultaneously. For example, some products may be in stock in St. Petersburg, but not in Moscow.

It’s vital for company marketers that customers see relevant prices in emails and do not receive recommendations for products that are out of stock in their city. Previously, contractor agency employees manually made mock-ups for each city: put down links to products and domains, and indicated phone numbers. It took about 7 hours to compile one mailing list for all cities: 6 hours for a marketer and 1 hour for the layout designer.

After integration with Mindbox, all work is done by a single marketer. He uses a universal letter template with a function called «regional data», which automatically puts down the domain, products, and essential links depending on a client’s city. To do this, we track 11 regional YML feeds hourly and update info if it changes. One newsletter takes up to three hours. Now, newsletter preparations take about 4 hours less to prepare.

Besides relevant information about products, local phone numbers, and other local data are automatically pulled into letters. Information about a client’s city is stored in a single client profile. If for some reason, a client’s city is missing in the database, then St. Petersburg is placed by default, because in this city, Petrovich has the most customers.

All links in a letter take the clients geographic zone into consideration.
All links in a letter take the client’s geographic zone into consideration.

Automated and personalized mass mailings

As of July 1, 2019, 55 triggered mechanics were launched on the project. Today we will tell you a bit more about the most interesting ones.

Two welcome chain options

About 80% of Petrovich clients are professional construction workers. The rest buy goods for themselves. There are welcome chains for each of these segments.

Letter
Letter

The welcome chain is different for private clients and repair masters.

The chain works better for technicians: the CTR is double

If the client’s profile does not contain info on the segment to which he or she belongs to, segmentation becomes the first communication step. The client receives an email with a built-in vote. To understand what positioning and button text will improve click rates, we conducted an AB-test.

Letter
Letter

Even a slight change in the letter layout affects the click rate

Option 2 showed the best result by click rate. The difference was only 0.4%, but on Petrovich’s scale, that amounts to hundreds of customers. The differences are statistically significant, with a confidence level of 95%.

If an individual indicates that he is a professional construction worker, he will receive a monthly business digest.

Digest
The digest contains three required blocks: a business literature book, “Petrovich” sales and construction market news

Reactivating clients with control group testing

For customers who have not yet made any order and have not visited the website within a month, we launched a reactivation mailing campaign. A control group was used to measure efficiency.

The target action for testing was the percentage of clicks to the site
The target action for testing was the percentage of clicks to the site

The hypothesis was correct. Customers who received the letter were 3 times more likely to go to the site than participants from the control group. The differences are statistically significant, with a confidence level of 95%.

Testing Next Best Offer on Machine Learning

Next Best Offer
Next Best Offer on machine learning, externally, doesn’t differ from regular recommendation letters

Using machine learning, we determine the best time to send a message with an offer to place an order. Unlike the usual Next Best Offer, which was sent to Petrovich clients 14 days after the order was placed, the new algorithm automatically calculates the optimal time when a client is most likely to place an order.

Made the omnichannel customer experience seamless

To decide on the purchase of decorative materials, many Petrovich customers prefer to have a look at the product first. So, they first check its availability on the website, then they come to the retail store, and then make an order online.

There are tablets in stores that allow customers to scan the products that they like. A shopping list that the client can forward to his email is formed. Such letters carry excellent indicators: open rate is 70%, click rate – 44%.

The use of tablets simplifies the purchase process and allows to save a list of selected products

Similar but more complex mechanics are implemented in the “AFIMALL City”, shopping center A Digital room is installed there, which allows one to create a design project based on the products presented in Petrovich and send it to an email address along with a shopping list. Such letters also come with good indicators: open rate of 69% and click rate of 29%. Clients use the data for an online order, but sometimes, they print the letter that they received and return with it to the retail store to finalize their purchase.

Digital room
Digital room

The digital room allows every customer to feel like a real designer and create a shopping list

Ekaterina Kirienko
Ekaterina Kirienko,
Email-marketer at Petrovich

«Mindbox is the aggregator of all our newsletters. Emails, Viber, and SMS messages are all sent through Mindbox. That means that without Mindbox, there would be no newsletters. Thanks to Mindbox, we are transitioning into an omnichannel business. I received tools that allow me to see live responses to the content that I create. For example, now I have a heat map that acts as a popularity indicator and shows who went where, spots with the most clicks, and what attracted the most attention. That’s very convenient because I understand which direction we need to move in and what we should think about as a company. My work has become much more convenient».

Used segmentations in triggered mailings

Petrovich marketers are now segmenting customers in the «abandoned category» triggered mechanic. Depending on the time of renovation work, 19 client categories were created. Each one came with its own type of individual email content. The result: the same quantity of sent letters resulted in double the number of clicks.

Used segmentations in triggered mailings
Used segmentations in triggered mailings

Used segmentations in triggered mailings

Conclusion and what’s to come

In a year and a half of working together, we increased revenue from direct marketing by 2.5x times, we created a single customer profile, combining all of the company’s services into one shared ecosystem. A complex communication scheme for commercial newsletters that utilizes the maximum number of channels for client interaction was developed.

Based on a single customer profile, we will continue to build omnichannel marketing. Using the collected data, we will be able to create even narrower customer segments. This will not only help personalize the newsletters as much as possible but also the website Additionally, we will implement new recommendation algorithms and show relevant pop-ups on the website.

This success story was prepared by

Andrey Medvedev

Andrey Medvedev,Project Manager

Philip Volnov

Philip Volnov,Lead Manager

Maria Bratchikova

Maria Bratchikova,Direct-communications
Manager at Petrovich

Marianna Lyubarova

Marianna Lyubarova,Editor
Subscribe to our quarterly news digest

Tell us a little about yourself

We’ll respond within 24 hours

Partnership request

Typically we’re answering within 24 hours