Reading time: 4 minutes | By Maarten Jansma, Account manager at Berkeley Bridge
Like many others, in the past, you have possibly used a voting tool to help you decide who to vote for. After completing a survey, users are shown how closely the parties political stances align with their own. This helps users make an informed, unbiased decision. A useful alternative for having to go through all election information yourselves, right?
Possibly without being fully aware of it, a voting tool is just one of the countless examples of tools that are based on a scoring model. Scoring models play an increasingly important role in our everyday lives, where every day a large number of decisions are made based on scoring models and a lot of advice is being given after completion of one.
The increasing pressure on organizations to make sound decisions faster and work more efficiently are driving organizations to make more use of scoring models. This trend triggered me to dedicate this blog on scoring models and how they come into being.
Scoring models come in different shapes and sizes. Some generic, others very specific. In short, you could describe a scoring model as follows; a model in which various variables are weighted in varying ways and result in a score. This score subsequently forms the basis for a conclusion, decision or advice.
In defining the possible outcomes organizations are not limited to numerical values, whereas the score can be translated to any format. Therefore scoring models can relate to different decision situations and answer a wide variety of questions, including:
Although scoring models come in all shapes and sizes, the common denominator of scoring models is that (specialist) knowledge is made available in such a way that less knowledgeable people are also able to draw the right conclusions.
In the past, Microsoft Excel was commonly used to share score models with end users. However, nowadays organizations mostly use decision tree based applications. These applications contain various (smart) questions or criteria to be answered until it generates a score. This method allows organizations to present complex information in a simple manner.
Example 1 – A risk analysis that uses a scoring model to determine to which risk category an event belongs.
Scoring models are also becoming increasingly automated. In that case, the model extracts all data (possibly from a database) that contributes to the scoring and automatically follows the defined steps of the decision-making process. This does not affect the outcome of the scoring model but does offer opportunities to further optimize the decision-making process.
A scoring model is the result of a scorecard card. A scorecard is a table in which all elements that influence the outcome are separated into individual characteristics, each with its own value. The various individual characteristics can, therefore, have varying influences on the overall assessment. To clarify this, please find below an extract of a possible scorecard.
Example 2 – A scorecard showing individual characteristics and their relative weightings.
Obviously, domain experts are responsible for specifying the scorecard. Their extensive knowledge and experience allows them to decide which elements actually influence the outcome and assign a score to each element based on its value. Additionally, it is important to realize domain experts are familiar with decision strategies of organizations and should conduct a periodic evaluation to make sure the scorecard still reflects the organizational strategy. Bearing in mind this shift of responsibilities and its evolving role, the domain experts’ role will possibly be a more strategic one in the future.
To emphasize how important it is that domain experts are involved in the process of creating a scorecard, think of a scoring model that is used to determine if there are any signs of symptoms of ADHD, and if so, to what extent. I am pretty confident you wouldn’t care too much about the outcome, if you’d find out the scorecard was created by your local baker.
When the scorecard is ready, the final step is to decide on the best way to make it available to colleagues. As mentioned earlier, this can be done through interactive or fully automated tree-based applications, for example.
With this blog, I have tried to give you a bit more insight into the phenomenon of scoring models. Although there is a lot more to discuss, I hope this has been a helpful start on gaining more knowledge about scoring models and realizing the usefulness of this method of sharing knowledge.
Curious about the many benefits of scoring models? In next week’s blog I will show you why an increasing number of organizations are using (and finding themselves more and more dependent on) scoring models and how it helps them realize their goals. I will illustrate the benefits with examples on how scoring models can support organizations in their decision making on the creditworthiness of (potential) customers.
Until next week!