Browse Prior Art Database

Learning to Rank Product Portfolio Elements via Expertsourcing Disclosure Number: IPCOM000244726D
Publication Date: 2016-Jan-06
Document File: 2 page(s) / 104K

Publishing Venue

The Prior Art Database


Large enterprises go through a massive exercise every year (fall/spring planning) where SMEs (Subject Matter Experts) or other business executives browse through the entire portfolio of products/services and revise the investment posture (development expense, SG&A expense, etc.) so as to push the top-line of the enterprise. Currently, in such an exercise, certain age-old managerial "best-practices" are used which are quite manual in nature. Human judgment/beliefs plays a big role in dictating the outcome of the exercise. Humans are limited by their cognitive capacity to crunch large amount of data. As a result, the correctness of human judgment/beliefs (that drive this important decision) is often limited. This may expose the portfolio at a risk that is hard to even quantify.

The idea here is to develop a tool that will aid human decision-making for such an exercise by means of eliciting and aggregating potentially diverse opinions of SMEs. The key challenge involved in developing such a tool is number of products/services which sometimes is so large that it is impossible for any SME to provide an assessment of the entire portfolio. Even if we restrict an expert to express his/her opinion on the most familiar portfolio elements, it could still be difficult for him/her to give an accurate ranking due to the large amount of data across multiple KPIs (Key Performance Indicators) such as expense, revenue, market share, no. of clients, etc.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 63% of the total text.

Page 01 of 2

Learning to Rank Product Portfolio Elements via Expertsourcing

Disclosed is a system and method to overcome the aforementioned challenges. A use case diagram of the proposed system is shown in the following figure.

Figure: A System to Learn Ranking of Product Portfolio Elements via Crowdsourcing

The proposed method used by the above system comprises of following steps

1. The proposed method first develops a list of all the relevant KPIs and collects the performance data for each of the portfolio element for each of these KPIs. These KPIs can be added/removed in a dynamic manner.

2. Next, multiple experts are invited to adjust the weights and performance ranges for the KPIs. Experts can modify these inputs through a UI and feel its impact on the portfolio ranking. This serves as a reinforcement signal for the expert. Experts continuously modify their opinions based on the reinforcement signals.

3. Based on the opinion of each expert, a separate ranking of the portfolio elements is generated.

4. All these rankings are finally aggregated to arrive at one single ranking, where aggregation can be done over any desired subset of the experts depending on the goal.

5. This final portfolio ranking can be supplied to any downstream use-case application pertaining to portfolio optimization. One use-case could be to decide which products should be iced. Another use-case could be to decide which products should be invested more in and which ones less.