Sergey Shivalin, a consultant for the creation of end-to-end analytics systems at UMSolution, spoke at Optimization 2019 how this analysis works on the example of a pharmaceutical company.
Reading time: 6 minutes
Tags: web analytics, marketing, online shopping
What is the article about?
ROPO Analytics helps you understand how online user behavior affects retail purchases. This relationship is of interest primarily to large retailers and online stores.
ROPO analytics will allow you to see what was previously hidden from you, change your data-driven marketing strategy and increase sales.
To determine the dependence of sales on Internet costs, you need a large system. Its development will take at least a year, this must be understood. It is best to enlist the support of management and not skimp on the team.
What is ROPO: who needs it and why?
ROPO (Research Online Purchase Offline) – users search online and buy offline.
For many companies, multichannel sales have been a godsend and a great boon. Now it’s more of a curse, because you have:
- points of sale;
- sites and promotion;
- mobile applications;
- messengers and social networks;
And everywhere a person can come into contact with you, get acquainted and interact. And elsewhere – buy. Do you know where your client came from? From Yandex.Maps, conferences, social networks? Most often there is no answer: well, a call came, for example, and came. What kind of call it was, how it came – such details are rather difficult to understand, and promotion also costs money. These are the problems that ROPO-sales analytics solves.
“The essence of ROPO sales: you marked the user with a tag on your resources, but on those resources where he made a purchase, you recognized him.”
Who needs ROPO analytics? Large retailers are interested in building ROPO sales reporting systems when they have their own online resources and points of sale. In addition, ROPO analytics is needed by online stores that have a showroom: they still have organic traffic from search engines ahead of other sources in terms of volume and importance: advertising, e-mail newsletters and social networks. Step-by-step work with ROPO analytics can look like this:
Collect and combine all transaction data for the required period: data online, from CRM.
Calculate the share of ROPO purchases in relation to purchases online / offline
Collect data about your ad spend.
Collect data on user behavior on the site, online and offline purchases. Look for the answer to the question – why the user did not complete the purchase on the site and went offline and draw conclusions by reviewing the marketing and advertising budgets.
Before conducting analytics, it is also important to get an answer to the question: why does your user leave the site to buy offline, or, even worse, to competitors, and does not complete the online purchase? Here we take into account several factors:
Site: it is simply inconvenient to make purchases on your site, this or that functionality does not work (for example, maps for choosing a delivery point are not loaded), the menu structure is not convenient for the user.
Trust: people want to make sure for themselves that the product is of high quality – to hold it in their hands, see it live, test it (if possible).
Time: they don’t want to wait for delivery – they want to buy here and now.
- Prices: are your prices online and offline very different from each other? Or, your segment of business involves the sale of very expensive (or very cheap) goods that people are used to buying on their own offline.
How did the ROPO analytics system work in the pharmaceutical business?
Sergey told how he built the ROPO system using one of his most ambitious projects as an example. There is a pharmaceutical business with such a problem: with many of its own resources, their main sales are in pharmacies (and there are several thousand of them throughout Russia). I would like to know: does what happens on the Internet somehow affect the final sales? With this question and a little interest of one of the owners of the company, the history of the creation of the system began.
Marketing in the pharmaceutical industry, like other industries, has a complexity: sales of many products are most often seasonal. There is a pronounced seasonal surge, into which the entire pharmaceutical business “throws” money, and there is no spending outside the season, especially on the Internet. In 2019, a lot has changed, a lot of useful content has appeared for representatives of the pharmaceutical business, but initially the Internet was perceived very badly by the “pharma”. You can learn how to effectively build work with the search space in the field of medicine in 2020 in this article. We share links to webinars on the topic of medicine:
Let’s return to the case and consider ROPO using the example of the seasonality of the pharmaceutical business.
The slide shows a diagram of the correlation between sales and customer costs for online advertising based on the existing marketing strategy. The seasonality of sales of sedatives is as follows: the peak is in December – since there is deep “alcoholism” and everyone’s nerves are at the limit, in January everything is declining. In the spring – also aggravation and recession again. Thus, the main sales were in January and March, and advertising budgets were allocated accordingly. The company did not notice that there is another, more important season – August-September.
Eventually ROPO-the report showed what was hidden
While competitors were investing in advertising in another potential season and increasing sales, the customer company passed “by” him, did not spend the budget on online advertising and lost billions of rubles. This was confirmed by the ROPO analysis. It turned out that the autumn period (late August-September) is a bright sales season for sedatives. The season when moms send their kids to school.
In short, the ROPO analysis revealed:
Additional growth points for online profits correlated with the seasonality of drug sales.
Serious mistakes in marketing strategy: the company did not properly build the budget, understand trends, seasonality and their correlation with sales.
Correlation between user behavior on the site and offline purchases.
- Large amount of competitor data: 30+ competitors and 1,600+ traffic sources. In addition, it turned out that the primary point of contact when choosing a drug is a search engine (Yandex).
ROPO analysis alone is not enough to track the effectiveness of a website and online advertising. It is necessary to constantly monitor trends, web analytics tools, techniques. We will discuss everything important to know about promotion in 2020 at Optimization 2020. In the section on Web analytics, experts and leading marketers of companies will present their successful cases, review effective tools and introduce you to working methods.
Features of ROPO analytics and working with ROPO reports
As a result of the analysis, a rather large ROPO report was generated. It includes standard reports with information about sales and sources, graphs with a breakdown by parameters, correlations of various metrics. Models for forecasting sales, traffic and seasonality were obtained. ROPO analysis made it possible not only to assess the market share of a business, but also to predict how budgets will affect changes in this share.
“Get ready for the fact that all work with data is not a very fast process. It will take a lot of time and patience. ”
A little about the system:
to parse SimilarWeb, Sergey’s team wrote a small neural network that parsed images (because SimilarWeb most often gives data in images). “Neuronka” on Pythone understood where the point was and recorded the information in text format;
parsing “Yandex.Wordstat”, parsing “Google Trends”, Semrush reports – all this is necessary for competitive intelligence. Analyzed the data of nearly three hundred competitors;
“Raw” counter data has standardized the work of Google Analytics according to its own resources, including even a strict order of id-goals and custom variables;
How to successfully introduce ROPO into your business? Work on bugs and tips.
Here’s what you need to know if you want to build a working system for accounting for ROPO clients:
Get ready to change the company’s business processes. This will take you most of the time, and this will be your main headache.
Prepare for fierce resistance from customer employees on all fronts. When people understand that a company’s business will become more transparent as a result of ROPO analysis, they will not want to provide you with data.
You will need an experienced team. A lot will depend on people: you will need analysts, developers, and those who will deal with business processes and visualization;
The completeness of the data is important. Make sure they are complete and formatted.
“Write documentation, describe architecture. Sooner or later it will save your life. Or a contract ”
Don’t prototype what you can’t implement. Sergei miscalculated when he made prototypes, presented them to the client and then prepared the contract. Some prototypes turned out to be easy only “on paper” and much more difficult to execute.
Don’t skimp on quality specialists! The client allocated insufficient budget – as a result, a team worked on the project without strong knowledge of building large-scale systems.
Do not forget to keep all the documentation, describe the entire architecture. Sooner or later it will save your life. Or a contract. On this project, Sergei seriously “pumped”, although he practically did not earn money due to fines.
The project lasted 1.5 years: when starting to work with such systems, one must understand that it cannot be done in less than a year. The return is about another year. The total budget is 4.8 million rubles. That is, the horizon for planning, developing and delivering such systems is approximately 2–3 years.
Sergey prepared a case for a pharmaceutical company several years ago, over the years, a lot has changed in his approach, and for sure a lot could be learned. You can learn more about Web analytics for E-com at the Optimization conference in 2020: at it, our speakers will talk about what new problems can now be solved and what can be useful from the experience of working in 2020.
– “What metric was used to confirm or deny the impact of offline on the online?”
Using parameter correlation. Sergey and his team checked a lot of parameters. For example, they found out which influences sales better – traffic from their own sites or traffic from different communities.
– “How did you connect traffic to communities with sales? Did you collect any data?
Information providers for the “pharma” helped with the database for analysis: they collect information about sales, traffic from communities, sites, etc. We analyzed data over the past 5 years, and then found a high correlation between traffic from the site, communities and sales.