FiNN FLARE case study: How to calculate the cost of delivery and ROMI considering purchased orders

Some of the orders that FiNN FLARE delivers at its own expense are not actually purchased by customers at the point of delivery. Therefore, the company decided to calculate how much they were actually spending on delivery and advertising in each region. Andrey Firsov, a contextual advertising specialist at FiNN FLARE, tells us about it.

Our clients
grow 22% faster

Grow faster by measuring what works best in your marketing

Analyze your marketing efficiency, find the growth areas, increase ROI

Get demo

About the company

FiNN FLARE is one of the largest fashion retailers in Russia and is an important player in the mid-segment. The retail chain includes 108 of its own stores and over 200 franchise stores in Russia, Belarus, Kazakhstan, and other CIS countries, as well as an online store. The company has been successfully operating and developing for more than 55 years.

In 2017, FiNN FLARE won the Narodnaya Marka Prize in the category “Chain of outdoor clothing stores,” taking first place in voting results among customers.

Task

We wanted to solve two main problems.

  1. Learn the actual ROMI of advertising campaigns considering purchased orders to more effectively allocate the budget across channels.
  2. Calculate shipping costs by region to find regions where it’s not profitable for the chain to deliver goods at our own expense and focus marketing efforts on other more promising regions.

Challenge

To attract traffic and increase conversions, we use many channels: contextual advertising, CPA (TrackAd), email, and others. Our advertising costs are calculated using different methodologies and stored in various services. In addition, we offer several shipping options, and in some cases, FiNN FLARE delivers orders at its own expense. Shipping costs depend on the region and the weight of the purchase. We also spend money on advertising in various regions, which increases the implementation cost. In order to regularly analyze all the costs of selling goods in different regions, we needed dashboards built on actual data, taking into account purchased orders.

When placing an online order, our customers have several payment options: pay on the website, pay in installments, pay when delivered by courier, or pay when picking up the order at a pick-up location. In the latter two cases, customers can try on merchandise and can choose to buy just part of an order. Therefore, the revenue that we see from orders created on the website is significantly higher than the actual revenue that we receive from purchased orders.

Data on delivered, filled, and purchased orders is also stored in different systems. To understand which products are purchased most often, whether advertising campaigns pay off, and whether delivery to certain regions is profitable, we needed to compile all our data into one report.

Solution

To set up advanced analytics, we turned to our partner CROC, a B2B integrator that has been working with clients in 42 countries for about 30 years. With the help of OWOX BI, the CROC team managed to collect data from different touchpoints and build dashboards that provided us with the missing information.

From checkout to purchase can take up to three weeks, so we use a report based on online data to make operational decisions on budget allocation:

However, we understand that not all orders placed on the website are purchased. In fact, this is a problem for most fashion retailers. Roughly speaking, a customer can pick items with a total cost of 50K rubles but only buy 15K rubles worth of merchandise. A report based on complete data shows the real picture: what percentage of goods are purchased and how much money is brought to us by specific channels.

For example, if you look at the ROMI of yandex/cpc, then in the first report, we see a figure of 419.46%, and in the second we see 115.97%. The difference is almost 3.5 times.

Therefore, at the end of the month, we look at the full data report and adjust the rates for the next month to consider redeemability.

The report on delivery parameters helps us determine in which regions there are more orders and cut off unprofitable regions:

Using a website and category report, we analyze which products are best sold on certain websites, in which websites we should invest, and which websites take time and don’t bring profit:

Since FiNN FLARE has many activities related to emails, we need to understand which campaign works best and encourages the user to complete the purchase. Thanks to the report on MindBox data, our marketers see all parts of the funnel from the moment an email was sent to the moment of purchase.

Insights obtained from this report:

  1. The conversion rate of emails sent before the actual purchase averages 0.2%.
  2. The largest percentage of conversions are made by customers who receive trigger mailings for product views and abandoned carts.
  3. Some users can open an email and make a purchase six months after receiving the email.

How we merge data for reports

Step 1. Set up raw data streaming using OWOX BI to simultaneously transfer data from the FiNN FLARE website to Google BigQuery, avoid data sampling, and build reports with all available parameters.

Step 2. Also with the help of OWOX BI, marketers set up automatic import of cost data from advertising services so data won’t be merged each time manually. This way, cost data is collected in one place and in a single format and is automatically updated regularly.

Data from Facebook, MyTarget, and Criteo is transmitted directly to Google BigQuery. Using the Google Analytics → BigQuery (Cost Data) streaming, it was also possible to get CPA cost data uploaded to GA by a third-party service. We upload Google Ads cost data into BigQuery using a native integration.

Step 3. Set up the upload of email campaign data using a custom integration by OWOX BI.

Step 4. With our own solution, configure the upload of transaction and delivery data from our internal systems to Google BigQuery.

Here’s a scheme of our data collection and movement:

Step 5. Set up OWOX BI attribution based on a funnel with two different conversion steps: placing an order on the website and purchasing the order. The result is two reports on the effectiveness of advertising campaigns: one report only on online data and one on complete data.

Step 6. We built several dashboards that answered important questions for us.

‘‘

I want to thank the OWOX team for their long and fruitful work. Issues and tasks were solved quickly. After integration, we began to spend less time collecting data for strategic decision-making. We are waiting for the next step with ROPO analysis integration.

Andrey Firsov,
Context Advertising Specialist ,
FiNN FLARE
‘‘

As an integrator, it was vital for us to organize coordinated work with the OWOX team to obtain a high-quality final result for the client. This was done thanks to a large number of prescribed processes and solutions already implemented at this point.

It’s important to note the constructive and prompt solution of unforeseen issues in close contact with the OWOX and FiNN FLARE teams, which made it possible to implement the project faster and more efficiently.

I am thankful to the OWOX and FiNN FLARE teams. After all, it’s always nice to work with professionals!

Vyacheslav Shuteev,
Product Analyst ,
CROC

Conclusions

As a result of a joint project to configure advanced analytics, the FiNN FLARE team was able to build reports that helped them answer previously unanswered questions. The calculation of real ROMI, considering purchased orders, allowed FiNN FLARE to efficiently distribute their budget among channels. And understanding how much it costs to deliver to each region helped them decide which regions were unprofitable to deliver to at their own expense.

The next step for the marketing team will be to finalize the upload of sales data in offline stores to build ROPO analysis. This will help to clearly track how online activities affect offline sales. Often, a campaign that seems ineffective online leads customers to offline stores.