Houston Analytics offers companies solutions and services for predictive analytics and knowledge management – in a brave and open-minded manner. The functionality of automated analytics solutions that are integrated into the customer’s processes and data systems is carefully confirmed before implementation. The customer can adopt proven analytics models right away.

Houston Analytics experts have the ability to quickly find the best solutions for information management from the outset of the customer’s business operations. In this way, Houston Analytics has provided several decision-makers with the opportunity to make decisions based on data and thus promote the company’s business.

The unique operating model, divided into three phases, functions regardless of the customer’s business operations and challenges. It is based on an agile CRISP-DM model (description on the reverse side), Houston Analytics’ extensive experience and best practices, and the technology of IBM, the market leader.

  1. Launch Pad – The experts of Houston Analytics’ Problem team examine the customer’s business needs together with the customer’s experts. At the same time, the Houston Analytics Solution team, in close cooperation with the customer’s representatives, examines what type of data and material can be used in the customer’s environment. Data entities are mapped and analysed from the perspective of business, and the opportunities presented by analytics are estimated.

After achieving a common understanding and vision, the entity is divided into decision-making points, and the type of customer’s data is surveyed: Can the answers to the customer’s process-specific data requirements be found within the available data? Is it necessary to enrich the data? As the end result of this phase, a concrete progress plan for information management is created for the customer.

  1. Rocket Launch – In the second phase, the quality of data in the customer’s environment is assessed. Useful data is identified and combined into the customer’s decision-making points from various interfaces. Houston Analytics experts carry out the physical preparation of data and evaluate how the data must be enriched to produce the optimal solution for the customer’s need.

This can consist of several physical analytics models constructed by the Solution team that will be validated before actual implementation: Did the model actually meet the customer’s needs?

The customer can start to use proven analytics models very quickly in their own operation.

  1. Mission Control – Analytics modelling has reached the target, that is, analytics models have been integrated into the customer’s processes and infrastructure as a seamlessly operating entity. A fully automated analytics environment can be implemented for the customer, for example, to control the company’s production and sales. The customer can monitor the selections made by the analytics if they wish to do so, such as the products the model proposes as products on offer and their prices. The customer does not need to have analytics know-how! Decision-makers can fully focus on their core competence and information management.

Significant benefits for customers

  • Comprehensive maintenance and development services at the service level (SLA) selected by the customer.
  • Expert services of predictive analytics, optimisation and knowledge management.
  • Continuous services also include the development of advanced analytics models (IBM SPSS Modeler) and data models (databases).
  • Customer’s data can be adopted efficiently and uploaded to a cloud.

Customers can use the services of Houston Analytics in a versatile manner to develop their own business. Companies are able to make smarter decisions with the help of precise, refined and enriched data.



The demand for data, and the volume of data needed, in controlling business operations is growing exponentially with digitalisation. Drawing a comprehensive picture of customer relations, the competitive situation, product portfolio, channel behaviour and trends requires robust expertise in research and analytics. As the business environment changes, the basic strategic questions – why, for whom, what, where, how and when – are more important than ever when analysing your situation and that of your competitors.

WHY answers the questions ’Why does the company exist?’, ‘What is its business idea and mission?’. It is a profound strategic question and clarifies the basic reason for the existence of the business. It states the longer term goals of the company and its purpose and role in society. Since no company exists in a vacuum, market and more detailed competitor analyses are great tools for clarifying purpose and place in the market.

The success of a business operation is always dependent on the customer, which is by FOR WHOM is an important question in any business. Who are the customers whose needs a company wants to fulfil, whom it wants to and must listen to, that is, what are the target groups of the company’s services or products? There may be several target groups, and they can vary by product group, product, area, brand or even channel. In an increasingly digital and fragmented market, segmentation based only on purchases or area is not enough. Understanding customers and engaging in dialogue with them must be at a deeper level, even 1to1. This will allow values, brand preferences, the customer’s share-of-wallet, purchase potential by area of need, and perhaps also channel behaviour to add significant value to the service provided to the company’s customers.

The question WHAT is tightly associated with why and for whom. What are the products and services offered to customers? When the economy was based on manufacturing, it was products and the management of the production process that were key. Today, both products and production processes area easily copied, which means that service has to be an integral part of competitiveness, at its very core. In the retail sector, product and service portfolios are traditionally set up based on size category while regional differences are not addressed. With optimising analytics, portfolios can be adjusted to regional demand and by location and concept.

The question WHERE links supply and customer demand and answers the question ‘Where can the customer be reached?’. What are the channels used to offer products and services to customers? Do they include brick and mortar stores, online stores or both? How do you manage the various channels? In addition to optimising excessive use of channels, analytics can be used to plan the locations and concepts of brick and mortar stores. What matters most is to view alternative solutions in the light of the demand potential in an area.

Today, when both analytics and varied marketing automation tools are available, the question HOW is often the way to competitiveness. It is not about what we want to communicate anymore but what, when and through which channel the customer wants to receive. How should the customer be served and how can the customer’s purchase path be supported as an individualised process that generates added value for the customer? How can we picture the customer’s needs, interests and priorities as comprehensively as possible? Fragmented demand and media use mean that a single product-price-media combination does not work with everyone. Analytics can help in finding the optimal supply-price-channel solution even at the individual level.

WHEN is a question that cannot be overemphasised in this age of digitalisation. In addition to conventional retailing seasons, providing supply at the right time is more important than ever before. Competition has become global and the supply cycle faster. But the biggest reason why timing has become so important is that supply has become more and more individualised. Customers are online and look for solutions to their needs here and now. Analytics, predictive models and tools that learn from customer behaviour are increasingly the answer to when. Marketing has become a service and its purpose is to make life easier for customers.

Price elasticity

Pricing has not lost its importance as a competitive tool, quite the contrary. Digitalisation has enhanced transparency, which has further increased the importance of pricing. Price elasticity indicates how the sale of a product changes when the price changes. When we have knowledge about the price elasticity and price images, optimisation can be used to determine the price customers are prepared to pay for the product. Target-oriented, profitable pricing integrally includes the monitoring of competitors’ prices which, combined with price optimization, enables improvements in profitability, even on a short term. Analyses make it easy to establish which products work best in each particular competitive situation.

Houston Analytics uses the IBM SPSS Modeler process to determine price elasticity. A linear regression model, implemented with an R node, is used for modelling. This eliminates the calculation of unnecessary statistical indicators, which significantly speeds up modelling – particularly with large data masses. As a result, we obtain product-specific price elasticity coefficients, scaled demand forecasts by price category, and sales forecasts for each store and day. Finally, the product-specific price elasticity coefficients are classified in four categories – inflexible, mid flexible, flexible and very flexible, which makes it easier to carry the results to practical decision-making and processes.




The retail value chain has changed during the past two decades: the trading sector has expanded its role from a distributor to an active operator which guides production through its own brands. On the other hand, the harshest change on the value chain has been caused by digitalization which has increased customers’ power by giving them access to the global supply that is now just a few clicks away. In the new business environment, the utilization of information provides the FMCG industry with a wide range of opportunities to improve customer understanding, processes and effective cooperation with the trading sector.

In the process towards information management, the first challenge is data: different business areas rely on their own sources and information has not been synchronized. As a rule, combining different internal and external data entities improves the explanatory and predictive power of the entity and provides a wide range of application opportunities.

Analytics serve decision-making above all. The viewpoint may be understanding the market: competitor positioning, understanding brands, pricing and trends. A set of its own consists of the management of customer dialogue and the related optimization of the media channel mix, launches and brand positioning. Analytics can also help all the way from purchasing guidance to the management of distribution channels and further monitoring. From the point of fruitful progress, it is essential to pose the right questions from the very beginning and to set objectives for the progress.



Internet has changed media behavior. Customers can look for anything, anytime, anywhere – according to their own preferences. Why would customers be any more satisfied with something for masses when they have a chance to take over the reins as individuals and construct their own media flows. This is the core of the online media business, but it also presents a challenge. Like in traditional stores, online success depends on the correct offering, that is, understanding customers’ online behaviour.

Customer loyalty can be increased by offering the correct content, with big data and analytics as keys. From the viewpoint of building customer loyalty, it is of key importance to understand customers’ media needs – viewing preferences and cross-viewing – but to also detect the first signals that indicate fading. The tools used are optimisation, segmentation and prediction models.

The basic analyses include utility models that are used to find out durations of customer relationships, program consumption, repurchases, seasonal fluctuations and user activity. This enables progress to more interesting analyses which can be used to predict the probability of program viewing or fading. The results obtained can be used when building content that better serves customers, more targeted offering and that way build a longer, more profitable customer relationship.


Predictive and Descriptive analytics
IBM SPSS Modeler Server
IBM SPSS Collaboration and Deployment Services

Optimization / Prescriptive analytics
IBM ILOG CPLEX Optimization Studio
IBM Decision Optimization Center

Reporting and Business intelligence
IBM Cognos

Business planning
IBM Cognos TM1

IBM Predictive Customer Intelligence
IBM Predictive Maintenance and Quality
IBM Counter Fraud Management