While paying so much attention to company’s online presence retailers start forgetting the old truth: 94% of retail sales are still rung up in physical stores and store location is still crucial for their success.
In busy cities with severe competition and extremely expensive real estate the cost of inappropriate store location can result too high for a company. That’s why we have decided to enable decision-makers with an innovative tool that will limelight the best secure way for their business to expand: we have built a mathematical model that pinpoints the best location for a next point of sale.
Traditionally the data used in a market research is limited to demographics, competitors’ location and social research data. While multiple social researches are costly and time-consuming for a company, a single research can not provide enough information to analyse historical, seasonal and other temporary factors.
The innovative approach by InData Labs provides a better solution because there is one more really powerful data source lying on its basis – social media data. Social media usually being used by retailers for communication purposes – in our research is used as an unlimited source of customer information. Our R&D team has produced a mathematical model that combines and analyses vast volumes of unstructured data to find areas with unsatisfied demand for different kinds of products and services.
Let’s have a look at some business-cases to better understand the key idea of our approach step by step.
Use Case for Telecom
Imagine a telecom company TeleTwee which decided to enter a highly-competitive Singaporean market. The company provides cutting-edge telecommunication technologies. As telecommunication equipment implies significant expenditure, TeleTwee can’t cover the whole territory of Singapore in a few months. Therefore they want to find the most populous regions to allocate the equipment assuming that it can accelerate return on investment. We know that the Singaporeans are active online in social media and blogs. Why shouldn’t we use social data to help TeleTwee to address the challenge and to find the most active zones within the city?
Our recipe is simple: we collect tweets from Singapore with geolocation enabled and then use our model to create informative maps illustrating most populous regions in Singapore based on tweets density. In the map below we can see how online activity is spread over Singapore territory.
The colored regions have a very simple and intuitive interpretation: warmer colors mean more benefits, since they indicate more popular areas. Our advice for TeleTwee is clear, the company should put its telecommunication equipment so that it covers these red hot spots and nearby regions. Here it is! TeleTwee can start conquering Singapore telecom market.
Use Case for Retail Chain
Let’s take a look at a more complex use case. Imagine a big retail chain TweeMart. TweeMart issue is much alike the previous example: its marketing team is looking for a place to locate a new store in Singapore. The challenge becomes more interesting as we should take into account competition from nearby shops. In this case, we need a bit more data besides tweets: we consider locations of other stores from TweeMart chain and competitors’ stores as well. We get information about other shops location from open sources. With this data our model can estimate how nearby stores influence demand in every single place in Singapore.
On the map the most attractive regions for a new store have orange color. Green and blue areas on the map are considered to be unattractive for expansion. The closer to the orange zone you are, the more chances you have to meet unsatisfied demand, if you place a new store there.
Use Case for Food Services
Sometimes it is helpful to take into consideration sentiments of tweets. Sentiment analysis uncovers insights about customers’ likes and dislikes. For example in our research we found out that the Singaporeans, in general, are more positive in their expressions: there were 15% more positive messages than negative in our dataset. Let’s consider another fictional example that demonstrates how sentiment analysis can be applied in practice and bring value to real businesses.
Lorenzo Tweemeto would like to open a new restaurant in Singapore and to show Singaporeans charming Italian cuisine. He knows little about Singapore public catering. As in the previous example he needs to find popular areas with minimum competition. Lorenzo Tweemeto has a secret idea that could raise his income: he believes that Italian cuisine could make people happier in the areas where they are not just starving but where they are looking for a better food quality and service. And it’s easy to find these places with our model. It can distinguish tweet topic and define its emotional tone. Analysing every tweet from Singapore and selecting messages about food and restaurants with negative emotional tone we help the enthusiastic Italian entrepreneur to pick the best location to launch its business.
Orange spots in the map are those places where a lot of people are socially active, where potential clients of a restaurant express demand for food and for better cuisine, while the number of competitors is lower than in the other regions of the city. With our model Lorenzo Tweemeto finds out where his clients are waiting for his services really fast, and he can start looking for premises for rent in a region that he likes most of all.
The mathematical model we used for these 3 simple cases is universal and can be successfully applied to any other type of products and services: grocery stores, banking offices or even more specific ones like a pet shop or a car wash.
Interested how the model works? Read our next blog post.
Using machine learning, AI and Big Data technologies InData Labs helps tech startups and enterprises explore new ways of leveraging data, implement highly complex and innovative projects, and build breakthrough AI products. Our core services include Data Strategy Consulting, Big Data Engineering, Data Science Consulting.
