Using AHP and PMLogic's DELIVER framework to make better project choices

Sep 25

6 min read

Updated: Sep 28

By James Bawtree | PMLogic B Corp

Using AHP and PMLogic's DELIVER framework to make better project choices

The real problem is rarely a lack of requirements. It is the absence of an agreed way to choose between them. The Analytic Hierarchy Process provides the structure; DELIVER turns that structure into a decision that can be implemented, tested and sustained.

The problem with calling everything essential

Most project teams have experienced the same conversation. The available budget will fund only part of the proposed scope, yet every stakeholder describes their requirements as ‘must have’ in other words essential. A workshop produces a list of must haves, should haves, could haves and won’t haves, but the labels do not help the sponsor decide which combination should be delivered first.

This is not simply a requirements problem. It is a governance problem. The organisation has not made its decision criteria, trade-offs and risk appetite explicit. In that environment, the loudest voice, the most senior stakeholder or the most recent issue due to our proximity bias can determine the outcome. The resulting scope may be politically acceptable in the room but difficult to defend later.

The Analytic Hierarchy Process, usually shortened to AHP, offers a more disciplined approach. Developed by Thomas Saaty, AHP breaks a complex decision into a hierarchy, compares criteria and alternatives in pairs, and converts those judgements into relative priorities. As Ricardo Vargas explains in his PMI paper, the method can combine tangible evidence with less tangible considerations such as strategic alignment, urgency, stakeholder commitment and risk.

AHP supports judgement rather than replacing it

AHP is sometimes presented as a mathematical answer to a subjective problem. That description goes too far. The numbers do not remove judgement because decision makers still choose the criteria and make the comparisons. What AHP does is expose those judgements, test their internal consistency and make the consequences visible.

The usual Saaty scale asks whether one criterion is equally, moderately, strongly, very strongly or extremely more important than another, represented by 1, 3, 5, 7 or 9. The even numbers can be used where a judgement sits between two points. The reverse comparison uses the reciprocal value. The completed comparison matrix produces a relative weight for each criterion and a consistency ratio. A ratio below 0.10 is commonly used as the threshold for an acceptably consistent set of comparisons.

That consistency test is valuable because it identifies contradictions. If a group says accessibility is more important than innovation, innovation is more important than cost, but cost is more important than accessibility, the conversation needs to continue. AHP does not conceal the inconsistency inside an average score.

Start with gates before priorities

Before ranking anything, separate genuine non-negotiables (must haves) from discretionary scope (should haves and could haves). A legal obligation, critical safety control, minimum cybersecurity safeguard or necessary technical dependency should normally be treated as a gate. If an option fails a gate, it should not remain in the contest simply because it performs well on other criteria.

This distinction prevents a dangerous outcome where a high score for customer appeal appears to compensate for a failure to meet an accessibility or privacy obligation. Once the mandatory baseline is established, AHP can help decide which remaining requirements create the greatest value within the available funding and capacity.

Bringing AHP into DELIVER

PMLogic's DELIVER framework takes the analysis beyond a one-off prioritisation workshop. It connects the choice to purpose, people, practice, platform and performance, then carries the decision through implementation, validation and learning.

Stage	AHP focus	What the team does
Discover	Frame the decision	Confirm the outcome sought, the budget and capacity constraint, the decision owner, stakeholders, mandatory gates and candidate requirements.
Examine	Build the hierarchy	Define a small set of distinct criteria, compare them in pairs and assess the requirements or capability options against them.
Learn	Challenge the evidence	Use lessons, operational data and stakeholder experience to test assumptions. Review contradictory judgements and the consistency ratio.
Implement	Fund a coherent release	Select the highest-value combination that fits the constraint, while accounting for cost, dependencies, sequencing and minimum viable capability.
Validate	Stress-test the choice	Run sensitivity analysis, check the proposed scope with users and operators, and confirm that the result remains credible when reasonable weights change.
Evaluate	Measure realised value	Compare actual adoption, service outcomes, cost and risk with the assumptions used in prioritisation. Identify gaps and unintended effects.
Reinforce	Make prioritisation routine	Refresh the model at agreed decision points, coach sponsors and teams in its use, and retain the rationale so future decisions begin with evidence.

Using the 5Ps to shape the criteria

A useful starting point is to test the criteria through PMLogic's 5Ps. This does not mean creating a criterion for every possible concern. Too many overlapping criteria can double count the same benefit and make the model harder to explain.

Purpose: How strongly does the requirement contribute to the intended public, customer or strategic outcome?
People: Who benefits, who is excluded, and what capability or change burden will the requirement create?
Practice: Does it strengthen governance, service delivery and repeatable ways of working?
Platform: Is it technically feasible, secure, supportable and compatible with the organisation's architecture?
Performance: What measurable value, reliability, risk reduction or operating improvement should it produce?

The final criteria should be specific to the decision. For one project, accessibility may sit under People. For another, it may be a mandatory gate. The hierarchy must reflect the organisation's strategy, obligations and values, rather than a generic template.

The problem with calling everything essential

A council meeting streaming service example

Consider a council developing a new service to stream public meetings. The team has proposed live high-definition video, captions, searchable recordings, agenda integration, speaker identification, automated translation, viewing analytics and social media clips. The available budget can deliver only about half of the proposed scope, and each requirement has been described as a must have.

During Discover, the sponsor clarifies the objective: increase transparent and accessible participation in council decision making while providing a service the council can operate reliably. Accessibility, privacy and cybersecurity obligations are established as gates. The team also identifies technical dependencies, including the need to capture and store a reliable recording before it can create searchable archives or clips.

During Examine, decision makers agree a concise set of criteria such as public value, accessibility, technical feasibility, operating sustainability and implementation cost. They compare the criteria in pairs, then assess the discretionary requirements. This is where the group may discover that it values reliable access and usable records more highly than promotional reach or advanced automation.

During Implement, the council does not simply choose the first four items on a ranked list. It selects a coherent release that fits the budget and respects dependencies. Live streaming, compliant captions, dependable recordings and agenda integration may form a viable first release, while automated translation, advanced analytics and social media automation remain candidates for a later release. The actual outcome depends on the council's evidence and judgements; the value lies in making that rationale traceable.

During Validate, the sponsor tests how sensitive the result is to changing the weights. If a small movement in one criterion completely changes the selected scope, the decision is fragile and may require more evidence. If the leading requirements remain stable across plausible scenarios, confidence in the decision increases.

Avoid turning the method into a spreadsheet contest

AHP becomes cumbersome when applied indiscriminately. Pairwise comparisons increase rapidly: 10 items require 45 comparisons, while 20 require 190. For a large backlog, group related requirements into capabilities, which is the ability to do something, use AHP to weight the decision criteria, and then apply an agreed evidence-based scoring approach to the alternatives. Reserve full pairwise comparison of alternatives for a manageable shortlist.

Five further disciplines keep the method credible:

Keep criteria independent enough to avoid counting the same value several times.
Record the evidence and reasoning behind significant comparisons, not only the number entered in the matrix.
Include the people who own the strategy, will use the outcome and must operate the solution.
Model capacity, cost, risk and dependencies after ranking. Priority is not the same as feasibility.
Revisit priorities when assumptions, funding, legislation or operating conditions change.

meeting AHP provides a structured approach

Better prioritisation creates better accountability

The strongest benefit of AHP is not a perfectly ordered list. It is a better-quality decision. Stakeholders can see why one criterion carries more weight than another, sponsors can explain why certain scope was deferred, and delivery teams receive a clearer basis for controlling change.

DELIVER extends that value by treating prioritisation as part of a management system. The organisation discovers what matters, examines the options, learns from evidence, implements a feasible choice, validates the assumptions, evaluates the outcomes and reinforces the capability. This closes the gap between selecting work and realising value from it.

When everything is called a must have, the answer is not a more forceful label. It is an agreed, transparent process for making trade-offs. AHP provides the decision logic. DELIVER makes that logic useful in practice.

How PMLogic can help

PMLogic helps executives and delivery teams establish practical portfolio governance, prioritise investments and requirements, strengthen sponsorship, and build the capability to make defensible decisions as conditions change. We have successfully implemented this approach for the NSW Department of Education and many other clients. Contact PMLogic to discuss a facilitated prioritisation workshop or a broader portfolio capability uplift using DELIVER.