Professional video analytics came to Russian retail

The first major introduction

Increasingly, Russian integrators receive requests from their customers for the installation of video analytics systems in retail. The first major retail chain to conduct a tender and lengthy tests to supply such a solution was M-Video. The main task in most stores now is counting visitors. However, the largest chain of stores paid special attention to such aspects as the accuracy of counting, the possibility of counting not only at the entrance, but also in an unstructured queue of inactive people, classifying personnel by clothes, determining the activity of staff, analyzing the behavior of visitors in the store. Due to the large number of stores , as well as with a variety of inputs in each of them, it was important to obtain generalized data on the number of logins from all inputs, to compare data obtained at certain intervals in different stores.

Counting visitors

Consider the features and benefits of video analytics for retail chains. Firstly, it’s counting customers . This function is performed by many counters without using video. If we compare the accuracy, then video analytics from a professional developer gives an accuracy of 95+ percent, which is more than enough for the customer. The accuracy of the meters varies from 70 to 95 percent, depending on the technology, operating conditions and the cost of the device. The main problems of the meters with limited functionality: low accuracy, inconvenience when processing data received in large retail chains.

Analysis of people in line

The next important functionality of video analytics is queue evaluations . Using this tool, you can estimate the number of people in the queue, the time they spend in it, get the average value for a day, a week, build graphs and charts that characterize customers. Why is the leader? Having such data, the manager optimizes the work schedule of the staff to reduce queues, and receives the cashier's efficiency coefficient (KPI). Some companies offer mobile notifications.(applications on iOs, Android, sms) about creating a queue for instant reaction of a senior cashier or store director. Using this tool, you can evaluate the quality of window dressing or mannequins in just one to two days (KPI of the marketing and merchandisers department). If the results do not suit the leader, then you can instantly change the layout of the goods and analyze the changes, based on the qualitative assessment obtained using video analytics. Now you don’t have to wait a few weeks for statistics on store receipts to accumulate. You save time, and properly designed windows and the correct layout of the new collection allows you to earn more money.

Customer Behavior Analysis

Another interesting feature of video analytics is the construction of a heat map . You get a card that shows where customers stop at your store and where they spend the most time. You can work with this card in a variety of ways. For example: identify the most expensive storefronts, place advertising information in hot places, more expensive products, draw attention to unpopular areas. Besides the heatmap, the administrator can also use the so-called map of trajectories . Using the map of trajectories, you can see the most popular travel routes for customers, identify unpopular areas of the store, take the necessary measures to attract people to the “cold” zones and evaluate how the behavior of customers has changed.

Analysis of decisions in the Russian market

Consider some of the solutions that are currently on the market. Firstly, it is the leading developer of video analytics company "Synesis". Specialized complex "Cassiopeia"Designed for retail needs. The user gets a kind of Google Analytics for offline stores with rich functionality. A high level of analytics is combined with a well-developed user interface, which allows you to get the most useful information for analysis and decision making. Foreign video analytics is represented on the Russian market by Verint, it is offered by several integrators. Verint produces high-level analytics, has successful experience in implementation in the European market. However, the disadvantages include the high cost of the solution, the lack of personalization and support on the Russian market.
Elvis-NeoTech also offers its solution. Its main feature is that video analytics does not use a computer or server and is performed in a computing box. Of the advantages can be noted a simple installation, low cost, promise an accurate calculation. Among the shortcomings are limited functionality (only counting), the ability to place only one signal line under one camera, the time-consuming update of software and video analytics algorithms, the lack of the ability to collect statistics centrally. Some companies are working in this direction, but have not yet developed a complete solution. For example, the Russian innovation company Satellite Innovation demonstrates the capabilities of its analytics.

Analysis of decisions in the Russian market

Cassiopeia video analytics complex
Manufacturer: Russia, Sinesis

• Pros: A wide range of modules (more than 5 different modules and more than 10 types of reports). Multi-server solution, works on slow communication channels.
• Cons: Lack of built-in ability to record video. But this flaw can be addressed with third-party free products such as iVideon.

Nextiva Video Analytics
Manufacturer: USA, Verint

• Pros: Adult, proven solution. Advanced reporting tools.
• Cons: There is no web interface, you need to install special software. You need a high-speed communication channel to a point of sale to receive a report. High price.

Statistics
Manufacturer: Russia, Elvis-NeoTek

• Pros: The easiest solution. Use of inexpensive cameras.
• Cons: Lack of centralized data collection and aggregation. It only supports counting on a single signal line without the ability to count people in the queue.

Macroscop and AxxonNext
Manufacturer: Russia, Satellite Innovation and ITV

• Pros: Multinational CCTV system.
• Cons: Only counting people. This solution was originally designed for security, not for business intelligence. There is no web interface with statistics and centralized data collection.

Scope of video analytics for business

The capabilities of video analytics are used not only in retail, but also in banks or, for example, in transport . The metro of the two largest cities of Russia is actively conducting tests among video analytics systems for security and video analytics for business, where the
Cassiopeia complex also participates in the final large-scale and multifunctional solution.

Tasks of the complex: passenger counting, queue assessment, passenger flow analysis.

User convenience and information security

Not only the accuracy of video analytics is very important, but also the convenience of the solution. The user should be comfortable receiving and analyzing data. No need to overwork the operator with unnecessary repetitive actions. It should be possible to view both standard reports and personal reports that this user needs. The ability to collect data according to various rules from one or several cameras. Also convenient access from anywhere to reports collected from all objects, for example, via the web interface, the ability to transfer data via an encrypted channel for user security is also important. It greatly simplifies the user’s ability to summarize and compare various graphs directly in the report without preloading the data. This allows you to create complex graphics for non-standard situations. For example,

Анализ с помощью видеоаналитики на практике

Consider the use of video analytics as an example of a clothing store. With all the necessary tools, managers will be able to identify weaknesses in their store. We begin the “diagnostics” using video analytics. First of all, you can estimate the number of customers who go to the store and compare with the number of people passing by (the rate of entry (ER - Entry Rate). If the percentage of visitors is small, you can work on the storefront and see how these data change. If with the number of visitors everything is in order, then we look further. We evaluate how customers move in the store. Do our expectations coincide with the real behavior of people. For example, we estimate how much time customers spend in a particular department, in front of a new collection or ed sales department. If you are not satisfied with the results, you can change the calculations "Images" additionally attract visitors in various ways. If at this stage everything is fine, then we look further. Next is the fitting room. You can assess the ratio of people who go into the fitting room with those who go to the cashier. If the difference is very large, it means that the buyer is not satisfied with the quality, style of clothing and something needs to be done. Be sure to monitor the absence of queues. This is a strong irritant for the buyer. The queue should not exceed 4-5 people, and the collected statistics will help to correctly distribute personnel between the ticket offices and the hall, depending on the time of day and day of the week. it means that the buyer is not satisfied with the quality, style of clothing and something needs to be done. Be sure to monitor the absence of queues. This is a strong irritant for the buyer. The queue should not exceed 4-5 people, and the collected statistics will help to correctly distribute personnel between the ticket offices and the hall, depending on the time of day and day of the week. it means that the buyer is not satisfied with the quality, style of clothing and something needs to be done. Be sure to monitor the absence of queues. This is a strong irritant for the buyer. The queue should not exceed 4-5 people, and the collected statistics will help to correctly distribute personnel between the ticket offices and the hall, depending on the time of day and day of the week.

The future of video analytics

Among the ten strategic technologies identified by Gartner for 2013,
at least three are strongly correlated with the tasks of video analytics for business: Strategic Big Data, Strategic Big Data, Actionable Analytics and Mobile Apps. Similar forecasts are made by IMS Research: Metadata as Big Data (Video Meta-data as “Big Data”) and Video Surveillance Goes Vertical.
Thus, the further development of the business infrastructure will necessarily include video analytics using all the capabilities. Particular tasks are collecting statistics, visualizing data for reports, watching videos online, predicting a queue, using a cloud service, and mobile notifications.
According to ABI Research, in 2013 the video analytics market will continue to grow rapidly, and by 2016 it will amount to $ 900 million.
As we can see, video analytics for businesses are also actively becoming interested in Russia. Big growth is expected in the Russian market for several reasons. Interest is not only from large retail participants (M-Video, Detsky Mir, Svyaznoy, MTS, RivGosh and others) but also from banks (Sberbank, Otkritie Bank) and small chain stores that seek to improve service and increase sales. Large projects are being launched with state financing in transport (Metro, Russian Railways, Airports) and in public institutions. Even large enterprises and factories strive to control the operation of the conveyor and jobs with the help of video analytics.