Case: Screening bids
The customer, a major Nordic manufacturer of process industry equipment, wanted to find for them more suitable requests for bids. Since they had experience about tens of risk factors they wanted to utilize this knowledge more efficiently by screening from their pool of hundreds of prospects those that best suited their strengths, and do it continuously at a lower risk level as presently. They also wanted a solution that their international sales organisation could use in everyday operations while updating and control responsibility would remain with the supply line. Such systems are not commonly available for manufacturers of capital goods. Since their global market potential was growing rapidly it would have been extremely risky to continue relying on just sales documentation and intuition. Therefore, BayesIT was called to design a bid screening system to suit their new needs. The development continues.
The customer did not have a screening system in use. They collected data about their prospects on a spreadsheet. Over the years it provided sufficient functionality as long as their offers were aimed at markets they knew well. With their increased activity on new market areas the risks increased since their knowledge of the new cultures was poor. A statistical selection approach was no longer adequate.
The customer experienced serious profitability variations especially when the buyer process was unclear and involved several decision makers. The problem was not only the increase of risks in execution of projects; the hit ratio was becoming so low that the cost of tendering became too expensive. The last point is extremely critical in a knowledge intensive environment where the capacity to produce good quality tenders was limited.
Bayes Information Technology Ltd., being a leading provider of Bayesian modelling, was called in to solve the problem. BayesIT designed a risk model consisting of three data sections; the data the company collected about the prospects, the history data of executed projects from their accounting and logistics systems and a new assessment process to collect so called tacit knowledge about the markets, customers, partners and subcontractors. The two first data sources, which relied on existing IT infrastructure, needed some further development to enable the knowledge model to identify the unsuitable prospects reliably.
A Bayesian Network model was created from this data. In the first phase the data was combined using a standard spreadsheet, and a new risk model was created when new data was available. Since the customer did not have an ERP in use, the customer had to manually combine the data from the various sources. The solution was designed so that it could easily utilize data from a future ERP as well as CRM system. The solution is easy to use, adding a new prospect, or new data about a project in realization, can be done in minutes, and does not require touching the system solution at all.
The Bayes Information Technology solution produces a short list of similar projects to the one under evaluation. The final feasibility judgement is done by the manager responsible for the bid. In case special engineering is involved, he communicates the risk factors to the front line managers for consideration before the joint bid is submitted. The rapid deployment of new data guarantees agility in the tendering phase. Cutting edge technology ensures that risk identification is accurate also when data in the early phases of tender preparation is very sparse.
With the bid screening system designed by Bayes Information Technology, the customers' engineering knowledge influences strongly the final bidding as well as take/leave decision done by the front line. As a consequence a truly managed decision considering up to tens of factors takes place jointly by front line and supply line managers. The replacement of intuition has saved not only a lot of costs but reduced the risk level significantly. Additionally, the co-operation between the supply line and the front line improved significantly.
The cost of the development project was approx. 100.000 euros and it took about two years to develop. The major part of costs was related to data collection development. At the time of order, the payback period was estimated to be 2 years. During the project some latent bad tenders were identified saving 0,1 - 0,2 million euros in tendering costs alone. Since it is very likely that some disaster projects were discarded as well, the system probably paid back in less than one year.