Data science in Sales - how analytics will help you earn more?

Data science in Sales - how analytics will help you earn more?

"Cheap, fast, good - choose only two options". Such a sentence you can read in on of the service points. The current customer has become a demanding person, aware and easily changing shopping habits.
On the other hand, companies that produce and sell goods operate in a competitive environment which is saturated in many areas of the market. Consider, for example, the telecommunications sector where the saturation of mobile telephony has exceeded 100% (according to the report of the Energy Regulatory Office of 2017) long time ago. So we have two different poles that need to be reconciled in some way.
Let's take a closer look at the perspective of the Customer and the Seller. Let's see how we can implement the latest technologies and analytics into this relationship, and who is required for this venture.

When buying a product or service, our customer has many purchase options. Traditional shop, where we can see the product live, internet - where we can use price comparison engines and choose one of the hundreds of stores. AliExpress Portal - where the price competition for local stores is absolute. What we decide on is determined by various factors - brand loyalty, price, marketing, Internet users’ opinions and friends’ recommendations.
However, the customers still have the same shopping behaviour. The first customers are those who regularly exchange their products (e.g. phones) for new ones. A very important group of Customers from the point of view of the seller - due to the regularity of purchases and selected products (mostly new products and therefore quite expensive). The second group consist of the Customers who, on impulse, make the decision to change the service provider or favourite store. For this group of Customers, a small lapse is often enough and they can never come back to us as a buyer. In the end, there are customers who make decisions in a long period of time - for them we need to prepare the substantive support and advice, so that the decision-making process ends in a purchase in our store.

Considering the above elements - the seller’s life isn’t easy. It is obvious that companies must earn. In most cases, there is additional pressure to improve the results - investors have their expectations. Traditional stores are losing customers for e-commerce. E-commerce - means competition, price war, proper logistics. Continuous development of the online store - introduction of new products (chatbots, bitcoins), innovations. Everyone wants to make his mark. And if the customer does not find himself on our site after a few seconds, he will give up and won’t come back probably.

In parallel with the evolution in the area of ​​the Customer and the Seller, fortunately IT technology advances - companies have the opportunity to collect data from many different places. The information systems they use to support their current work are a mine of knowledge that can be enriched with additional external data allowing even more comprehensive look on the information.
Of course, there are a lot of data, they appear in different cycles (often online) and are stored in different ways, but also here technology makes us deal with this area quite well. Technology providers have discovered the potential hidden in their data several years ago and the Big Data technology has been created. Its development continues untill now. Thanks to such an approach, the volume of data and processing speed cease to be a problem, and strong competition allows access to these solutions at a relatively low price.

The development of Big Data technology goes inseparably hand in hand with the analytical solutions that allow us to extract important information from our data collection and then continue to work with it. Both, in the area of ​​ongoing analysis and reporting (traditional Business Intelligence approach), as well as in the area of ​​predictive analysis - which helps us  to find answers to questions related to various aspects of our business in the future. Predictive analysis gives measurable results when it knows about the prevailing trends in the business of our organisation in the past. That is why the aspect of collecting historical data is so important. The more good quality historical data we have, the better predictive models results there are.
Once we build a place for our data (data warehouses are the most popular solutions), we need two more things: an analytical task and people who will build analytical models. Predictive analysis is a field closely related to statistics and mathematics - the creation of the appropriate models requires strong grounds in this area. In addition, there is knowledge of relevant information technologies, but about it you will know soon.
An analytical task is a model that we want to build to analyse a specific business area. It may be a sales analysis and an attempt to forecast the sales trend in the coming months. These can be models related to the rise in sales, allowing to forecast products or services that our clients want to buy, based on the historical behaviour of the similar customer groups. The analytical task must be linked to the real business need and must be measurable in the context of verification of the results obtained by the model.
Remember, if we skillfully and fairly detail our clients into smaller groups of behaviour, they will be often very similar. Therefore, learning how this behaviour looked like, we are able to prepare analytical models and use their knowledge during the current purchase transactions - classify the customer properly and then lead him through the shopping path.

Predictive modelling is a work combining competences from three areas - mathematical science, technology and the ability to work with a business client (communication and analytical thinking). Companies where analytics is an important element of business support are increasingly building their own Data Science teams. The big advantage of such an approach is the fact that analysts have extensive business knowledge in the field of the company's functioning what gives a proper result into the solutions they build. However, most companies use the services offered by IT providers, where also dedicated Data Science teams are set up. In this case the advantage is the project approach (which results into a one-off project cost) and the experience of consultants working with different clients and data.
Analytical projects compared to other IT projects are relatively short and that is why the cost of implementation is quite low. The average duration of such a project of a specific analytical task implementation by BitPeak company, is less than 3 months. The experience of BitPeak in the field of analytical projects shows that the most difficult is to start and to find the first real topic for analysis and people willing to get involved in the venture. However, if we start and what is more, the predictive model is effective and measurably improves company’s results, then there will be many others topics to consult. Predictive analytics can be successfully used in many places of business processes functioning in the enterprise.

How to start
We need 3 elements - business needs - which I hope that after reading this article were managed to create, technology and implementation partner. If the organisation has a problem with the analytical task identification, BitPeak as a part of its methodology offers a cooperative consulting workshop during which our consultants will help to gather the requirements and to establish a real business need for which we can choose the analytics.
In the field of technology, there are Enterprise class solutions (such as IBM SPSS) dedicated to enterprises that are mature in the field of analytics, that have their own teams and whose business uses the results provided by the analyst intensively. However, for smaller solutions, it is worth to be interested in the environment available in the cloud (IBM Data Science Experience) or create a solution based on the open-source technologies that in the analytics or Big Data area have commanded a significant part of the market (HortonWorks, Cloudera or R language programming ).

If you decide to trust BitPeak while choosing an implementation partner, I invite you to contact me. Our company consistently invests and develops Data Science practice as one of the strategic areas of activity.
About author
Sebastian Kaminski - Head of Information Managment & Analytics in BitPeak. Previously managed BI & Analytics team in Atos Poland and Involved in Business Intelligence project for more than 15 years in different roles. Lecturer at Kielce University.