OFD, Retail

Product analytics

We have over 15 years of experience in generating all types of analytical reports:

• We make Excel-files (unloadings) with sales volume for target product and its price dynamics.
• We build dynamic analytical dashboards in BI and web interfaces that provide real-time analytics.
• We build buyer profiling models based on preferences and consumer behavior.
• We conduct a statistical analysis of sales dependence on external factors.
• We conduct Ad-hoc studies on customer request.
• We do demand forecasting and predictive modeling.

Data labeling and categorization

We have an experience of creating an automatic categorizer of goods, which is implemented into the data processing system of OFD operator and allows data separation not at the client’s request, but constantly. This greatly speeds up the output of the markup result.

We markup a large amount of data of hundred million lines, because we have:

• a set of modules with data sales categorization,
• an internal system for preparing, managing and balancing training samples,
• staff of specially trained markers,
• distributed computing infrastructure.

We worked with the categories FMCG food/nonfood, Beauty, Horeca, Farma, Sport goods, Pet products, services, OFD data of traditional and modern trade, e-com, banking data, telecom and media data, text and other data.

*Fiscal Data Operator

Segmentation of the trading space

We have developed a segmentation model that automatically determines which of the business segments a particular sales outlet (cash register) belongs to.

 Many parameters are used to define a segment, such as:

• econometric indicators of the point of sale,
• semantics of sales in checks,
• OKVEDs* of outlets.

We implement this system into the OFD data processing circuit.

*Russian Classification of Economic Activities or Russian National Classifier of Types of Economic Activity

Customer segmentation

Our customer clustering and segmentation models allow us to divide customers into groups with similar preferences and consumer behavior.

The model performs segmentation based on sales data and loyalty system data. Then, using the Look-a-like module, the built profile of the target audience can be expanded to the volumes necessary for placing advertising campaigns and marketing campaigns.

Data cleaning and calibration

Raw data is most often unsuitable for qualitative analysis, because data entry errors at the checkout, differences in the culture of data entry into a check in various retail chains, wholesale trade, price trends, seasonality, etc. can distort the result of the analysis, so the data must be prepared in advance.

 Our cleaning algorithms effectively remove outliers and data errors. The use of calibration and data weighting techniques allows obtaining representative samples at the level of the country, regions, strata, cities, business segments, product categories.

We are able to create ready-to-operate dynamic data quality management systems and integrate these systems into the Customer’s circuit.

Data visualization

We present large amounts of data with complex relationships in the form of interactive dashboards for quick situation analysis and management decision making. And the dashboarding department develops visual, immersive BI dashboards and heat maps that show market dynamics in real time.

Extensive experience in OFD, retail, digital, telecom, media and other industries allows us to recommend the most useful data sections to our clients.

We also create interactive websites and applications, we can integrate the application into the client’s CRM/ERP system* or set up email notifications about significant market events.

*customer relationship management system

Employee Profile Record

Building predictive models

Predictive models allow you to make a forecast about a future event with a certain degree of probability, find out what sales will be, how the buyer will behave and what will be in demand on a certain forecast horizon.

We have experience in building predictive models for price and demand.

We can offer a system for evaluating the effectiveness of promotions, namely analyze the effectiveness of promotional activities and marketing campaigns of a similar type in history and form an assessment of the effect for the planned promotion.

Recommender systems

Competitor analysis

Analysis of SKU distribution, competitors, category structure is the first mandatory step in developing a product marketing strategy. A correct assessment of the company’s competitive environment allows you to create a sustainable competitive advantage of the product, choose the right communication channels and reduce operational risks. We can visually build a tree of competitors and a related products graph.

Brand switching

We analyze brand-to-brand switching:

• ind reasons for switching and identify socio-demographic characteristics of switched clients,
• display the top competitor products by switching stores, regions and countries,
• display the switching shares in money and the quantity of goods.

We can show all of this in dynamics, as well as display it in the form of standard reports, interactive BI dashboards, web interfaces and applications.

Brand building

Brand differentiation from competitors is marketing campaigns or brand positioning in such a way as to stand out from the rest of the companies and attract the attention of the consumer.

We can conduct a classic survey or organoleptic testing on competitor brands, enrich sales data, form a publication or advertising strategy and implement it using the best practices in advertising.

Statistical analysis

We help to obtain additional information that is not obvious at a simple glance. For example, statistically significant dependences of sales volume on external factors (air temperature), which we did for a popular brand of sweet carbonated drinks. Or examples of beer brands whose sales growth occurred due to an increase in the frequency of purchases.

Calculation of elasticity of demand

Calculate the price elasticity of demand for specific SKUs based on historical sales data. It will show the percentage change in demand for a good or service as a result of a change in its price.

Planograms

We will make a visual display of the optimal display of goods on the shelf in order to increase sales of this category. It is built on the basis structure analysis of the category and retail space, geographical and seasonal characteristics of the product, as well as on the basis of knowledge of consumer behavior and preferences.

Econometric analysis of business segments

We implement an econometric analysis of business segments and calculate the characteristics of retail outlets:

• average check,
• sales volume
• number of checks
• share of free price sales,
• number of positions in the receipt,
• assortment (number of categories, number of SKUs in total and in each category).

We do clustering within the segment according to any criteria, which allows us to understand the structure of the trading space.

Product distribution analysis

We analyze the number of outlets with goods in the total number of outlets, analyze competitors, market share, by brands, city strata and regions.

Such an analysis makes it possible to assess the potential of the brand in regional outlets, increase market share and sales, give an understanding of the target profiles of outlets, and cut off outlets with no prospects for contact. And also to propose a strategy for working with outlets, to build a model for predicting the potential of an outlet based on the structure of its sales.

Structure analysis of commodity category

We calculate the dynamic and static indicators of categories by regions and city strata: price, sales volume, average bill. This allows us to understand the structure of the trading space to build a strategy for launching a new product on the market.

Penetration of goods into category

We determine the share of checks with SKUs in the category, brand market shares, search for related products (product pairs). This gives an understanding of the cause-and-effect relationships of changes in market volume, sales of competitive brands and relationships with sales of competitors, reveals points of growth for business segments.

Price dynamics of a product or commodity category

Tracking the general price trend of a category, comparing the target SKU with the dynamics of competitive brands in terms of prices, comparison with previous periods - help to get a general understanding of the price market.

Dynamics of sales volume changes

Tracking the general trend of the category, comparing the target SKU with the dynamics of competitive brands in terms of sales volume, comparison with previous periods – allows us to get a general understanding of the market.

Shop-centric panels

We have experience in creating representative panels (information sampling/access acceptor) of retail outlets for analytical research in conditions of high rotation of research objects.

People-centric panels

We have experience in segmenting and clustering consumers according to preferences and purchasing behavior and creating representative panels (information sampling/access acceptor) aimed at consumer research.

Text analytics

Helps to identify meanings, essences, relationships from unstructured texts. We automate the processes of extracting and structuring important information from text data.

Using text analytics, we can select keywords based on customer segmentation by preferences and consumer behavior for brand advertising.