How predictive analytics improved Forbo's sales results

You likely walk on a Forbo floor quite frequently. Millions of the company’s floors have been installed in residential houses and offices around the world. Forbo Flooring produces floor coverings for the project and consumer markets and special flooring for spaces like medical operating rooms, server rooms, shops, dental practices and schools. With a 65% market share, Forbo is the global leader in linoleum floors. 

The situation
Sales opportunities with the highest potential are selected based on a ‘gut feeling’. Unreliable prediction of the sales results.

The solution
Calculating reliable chances of success using predictive models based on big data.

The result
Potential of 1-3% improvement in turnover and 10% cost savings.

Making the smart choice

Roughly 70% of the floors that are sold are for professional use. René van Vliet, Sales & Marketing Controller at Forbo Flooring, explains the challenges for this market: ‘each year, more than 3,200 projects that are relevant for us, start in the Netherlands. Several market parties are usually involved per project. We can’t possibly follow all the opportunities in detail, and therefore must choose which projects and involved market parties to approach. In the past, we did this based on feeling, but we were convinced that better decisions could be made based on smart information.’

Rene van VlietWe can make better decisions and better predictions, but above all, we can conduct better price negotiations.
René van Vliet Sales & Marketing Controller, the Netherlands, Forbo Flooring

Classification

In 2014, René’s department, therefore, started collecting as much relevant information on projects and the involved parties, as possible. How many projects does an architect have going? How large is the architectural firm and is it in good financial condition? What is the recommended floor surface for the project in question? Information, obtained from the Chamber of Commerce and Dun & Bradstreet, was added for more than 4,300 architectural firms. Using this information, all firms were classified based on attractiveness and potential for Forbo.

‘Many businesses look at historical information. How did we perform in the previous period? Our goal with this information, was to be able to look ahead. More accurately assigning priorities to opportunities and being able to better predict the commercial results,’ explains René.

Predictive model

In 2015, together with Kadenza, Forbo explored the possibilities of converting the collected data into meaningful information. ‘We were very curious to know if we could predict the likelihood that we could convert a project and what the decisive factors were,’ explains René. ‘It soon became clear that more data was needed if we were to build reliable predictive models. We added information on quotations, invoices, sample requests and sales activities and Kadenza developed a predictive model with IBM SPSS.’

More correlations

With IBM Watson Analytics, correlations are subsequently sought in the immense volumes of data. ‘The great thing about Watson Analytics is that you can simply ask questions without the need for IT knowledge. For example, we checked if there are differences per account manager and whether these differences had anything to do with the sales region.’ This provided a great deal of insight on which factors are decisive in a project’s potential for success. The chance of success can now also be calculated per opportunity, which means the prediction of future sales results becomes significantly more reliable.

Amarins van de Voorde KadenzaWith predictive analytics, you exponentially increase the value of your organisation’s data, resulting in a significant improvement in your operating results.
Amarins van de Voorde Consultant, Kadenza

Operationalisation

René will soon take the next important step, namely, operationalisation of the insights obtained. ‘Now that we know what determines a project’s chance of success, we will ensure that the chance is updated for every opportunity, based on the current situation, so we know what our chances are in every selling stage of a project, what the priority is and on which party we should focus.’

Results

Although the approach’s exact operating results will only be visible in a while, René already has high expectations. ‘We expect to make better choices and better predictions. Moreover, we’ll be able to conduct far better price negotiations. An overall price improvement of 1% would already be amazing, but I think that 3% is feasible for many of the projects. By spending our time on the right things, I also think we’ll be able to reach a cost reduction of 10% in time. That would be an amazing result.’

Pleasant collaboration

Such results will certainly motivate Forbo to get more predictive value out of the available data. This also applies for other business units. Would René choose Kadenza again? ‘Yes, certainly. Kadenza truly joined us in the thinking process for this project. The consultants delved deep into Forbo and our specific request, to then make the transition to the best and most suitable solution.’