Follow

Analytics Strategies for Different Decision Types

Have more questions? Submit a request.

Think of Decision Lens as a benefit-cost analysis tool. This is a concept that almost everyone easily understands. In Decision Lens, you build your "benefit" model with your criteria and rating scales, describing the specific sources of value stakeholders seek from their investments. If appropriate, you can build a "cost" model in Resource Allocation, and complete the benefit-cost analysis to inform a variety of decision types.

Unlike traditional benefit-cost analysis, which requires expressing costs and benefits in financial terms, Decision Lens allows for relative measures of benefit and cost. This greatly expands its applicability. Effective analytics begins with understanding the decision objectives and structuring data around them. Think of analytics as a design challenge: what pieces of data can we work with to help stakeholders derive insights from their portfolio without moving mountains? What specific metrics and visualizations will speak directly to stakeholder questions without overwhelming their senses with data? What existing and commonly-accepted organizational metrics can be applied to accelerate the decision process and quickly arrive at insights?

The first step in developing your analytics strategy is to gain a clear understanding of the type of decision, or use case. There are several basic types of use cases, and each one implies specific data, scenarios and visualizations to enable the best possible insights for stakeholders.

Simple Prioritization

A simple prioritization produces a 1-N ranking of alternatives by Priority Value score based on criteria, ratings, and priority weights. The Priority Value score indicates the relative amount of benefit or value offered by each alternative to stakeholders. This outcome is visualized in Sensitivity Analysis

A simple prioritization identifies the best option among competing options where the cost of each option is commensurate, or cost is not a key factor. Typical use cases include candidate hiring decisions, vendor selection, strategic prioritization of product or market opportunities, or long-term strategic planning.

A few things the Analyst should consider in planning a simple prioritization:

  • Size of the portfolio: Bigger portfolios (>50 alternatives) are not ideal for 1-N rankings without some type of segmentation variable to make them more digestible for stakeholders. Segmentation can occur through alternative categories, allowing stakeholders to evaluate peer groups of alternatives.
  • Stakeholder groups: Segmentation by stakeholder group is also a good option for larger portfolios. You can create alternative categories for stakeholder groups, and apply them in a Bubble Chart to show peer groupings by color. See the article improve your simple prioritization for tips on this technique.
  • Equity of outcomes: In some cases, a simple prioritization may be constrained by an organizational need for equity of outcomes across a certain dimension, such as departments or functional areas or geographic areas. This is easily achieved using alternative categories and selecting within the filtered categories.

As noted above, a simple prioritization carries the implicit assumption that cost is not a key factor nor a differentiator in the decision. However, if cost data is available and might be a differentiator (especially in larger portfolios), there are several ways you can improve your simple prioritization using cost data to enhance your rankings.

 

Budget Allocation

A budget allocation decision prioritizes alternatives in order to derive maximal value from a limited, or constrained, set of budget resources. Unlike most simple prioritization decisions, a budget allocation decision operates under the assumption of resource scarcity, and aims to find a value-maximizing combination of projects to fund without over committing resources.

A budget allocation decision will require the use of Resource Balancer, and with it data sufficient to build resource pools and cost profiles to build and run a budget optimization scenario.

Budget allocation and optimization scenarios can get pretty complex pretty quickly, but there are some simple planning decisions that can help this decision more manageable. Here are the key things to consider:

  • How important is "bang for the buck?" The default optimization approach in Resource Balancer is maximizing Value Return on Investment, or "bang for the buck." However, the importance of resource efficiency in the budget is something that should be clarified with stakeholders before building scenarios. If stakeholders want to focus more on business value than overall resource efficiency, Analysts can adopt a different approach in designing scenarios and visualizations to help stakeholders achieve this outcome.
  • What is the distribution of alternative costs? Take some time to assess the distribution, or spread, of alternative costs (creating a histogram of the cost data in Excel can be useful here). A large spread of alternative costs (by a factor of 10X or more, when comparing the highest vs. lowest cost alternatives) will generally favor the lower cost items during resource allocation. If the portfolio has a relatively small number of very high cost alternatives (we call them "whoppers") competing with a large number of low cost alternatives, have a discussion with the stakeholder team about the criticality of funding the "whopper" projects. These would need to be force funded in a more competitive scenario in order to include them in the recommended portfolio.
  • Remember, the goal is to narrow down the array of choices. Decision Lens is not designed to replace human judgment in decision-making. It is designed to improve it. Large arrays of choices overwhelm human cognitive capabilities very quickly, so Decision Lens can be very powerful in narrowing down the array of choices to a relative few so that human judgment can be applied. Ask a group of executives to prioritize 150 IT projects for budgeting and their faces may go blank. Ask the same group to compare a dozen or so projects with the aim to identify the best choices using comparative metrics, and they will have an opinion. Decision Lens excels by focusing stakeholders on a relative few remaining choices after simplifying large portfolios.

Resource Leveling and Scheduling

While budgeting scenarios are almost always constructed using financial resources, a related decision type might involve leveling and scheduling physical resources tied to projects funded in a budget allocation decision. This is common in IT. The idea here is that you can separate the project selection decision from the project scheduling decision. They are closely related, but in an environment as dynamic as IT, it is better to separate them using different scenarios (in the same model). Prioritize and select projects for funding first, then schedule them against your most bottlenecked resources.

  • Time period: In annual budget allocation decisions, there is rarely a need for multiple time periods (this changes for long term capital planning). However, resource scheduling involving people or other physical resources typically requires monthly or quarterly periods to balance availability vs. demand. While executives are not interested in the intricacies of scheduling, managers are very concerned about this issue. Thus separating the project funding decision scenario (using currency) from the scheduling scenario (using FTEs or other pools) is a good design approach.
  • For multiple resource pools, focus on the most constraining resource(s) in your scenario. IT organizations typically have dozens of types of technical resources required to deliver projects. Yet many of them likely do not represent true resource bottlenecks for the organization. The most critical constraints often involve only a relative few roles, such as (or example) project managers, database engineers, and UI developers, are in shortest supply. Building the allocation scenario around these relative few resources that represent the most constraining shortages will keep scenarios more manageable without oversimplifying the results.
Was this article helpful?
2 out of 2 found this helpful

Comments

Powered by Zendesk