Why Consulting Projects Fail
Consulting projects almost never fail because of bad analysis. They fail because of vague scope, missing governance, slow decisions, or a team that discovers in week six that the client expected something different in week one. The technical work is usually fine; the management discipline around it is not.
Project management for consultants is a different discipline from project management in industry. The deliverables are documents and decisions, not buildings or software releases. The team is often virtual, multi-firm, and changes mid-engagement. The client is paying premium rates and expects a level of crispness that internal project management does not have to provide. This guide is the playbook we use to run engagements that finish on time, on budget, and on the original scope.
The Five Phases
Every consulting engagement, regardless of sector or methodology, passes through five phases. Skip one and quality drops in a predictable way.
1. Inception (Weeks 0–2)
The inception phase exists to convert a signed contract into a working plan that both sides agree on. Done well, it surfaces every misalignment between the proposal and the client's actual expectations while there is still time to fix them cheaply.
Inception deliverables, in order of importance:
- Refined methodology. What you actually intend to do, written in language the client uses, not consulting jargon.
- Workplan with milestones. Two-week granularity is right. One-week is too detailed to maintain; monthly is too coarse to manage.
- Stakeholder map. Named individuals, their decision-making power, their stated interest, and how often you will engage them.
- Risk register. Five to ten material risks with owner, likelihood, impact, and mitigation. More than ten and you have not prioritised; fewer than five and you are not being honest.
- Governance arrangement. Who chairs the steering committee, who sits on it, how often it meets, and what decisions it owns versus what decisions the engagement manager owns.
The inception report is signed off by the steering committee chair before any substantive analytical work begins. Without that sign-off you are working on assumptions that may not hold.
2. Mobilisation (Weeks 2–4)
Mobilisation is the unglamorous phase where the team gets access to data, schedules interviews, sets up the working space (shared drives, project management tool, communication channels), and books the diaries for the next 12 weeks. Most engagement managers under-invest here because mobilisation does not feel like progress; it is, and skipping it pushes pain into the analysis phase.
A mobilisation checklist that has saved a lot of weeks:
- Data request list issued to client with a named owner and a deadline
- Interview programme agreed: who, when, by whom, for how long
- NDA exchanged with any third party that will see client material
- Shared workspace stood up with folder structure that matches the workplan
- Weekly stand-up calendar invitation issued to the engagement team
- Bi-weekly client check-in calendar invitation issued
- Steering committee dates booked for the duration of the engagement
3. Analysis and Synthesis (Weeks 4–12)
This is the phase clients are paying for, and it is also the phase most likely to drift. The discipline that holds it together is the storyline-first approach: write the table of contents for the final deliverable in week four, and update it every week. The storyline forces a hypothesis you can pressure-test, identifies the analyses you actually need, and stops the team from generating slides that will be cut.
The single most common failure mode is over-analysis. A consulting engagement that produces 200 slides has usually not earned its fee; it has demonstrated activity. The mid-term review (typically week six or seven) exists to force pruning — every workstream lead presents what they would cut from their section if they had to cut a third.
4. Delivery (Weeks 12–14)
Delivery is more than the final presentation. It is the orchestrated handover from the consulting team to the client, designed so that the client can act on the work after you leave. A common mistake is to treat the final deck as the deliverable; in reality the deliverables are the deck, the supporting analysis, the implementation roadmap, and the answers to questions the client will ask three weeks after you have gone.
A complete delivery package:
- Final report or deck (the headline output)
- One-page executive summary
- Detailed appendices with underlying analysis
- Implementation roadmap with named owners on the client side
- Data files in a format the client can re-run
- A 60-minute Q&A session two weeks after handover
5. Close-out (Week 14+)
Close-out is where you protect the next engagement. The internal lessons-learned session, the client satisfaction conversation, the case study write-up for marketing use (with client permission), and the formal handover of intellectual property all live here. Most firms skip close-out and lose the compounding value that comes from running each engagement as a repeatable system.
Governance That Works
Governance fails when it becomes performative. A steering committee that meets monthly, listens to a status update, asks no decisions, and disbands is theatre. A steering committee that meets fortnightly, sees three decisions presented as A/B options with the team's recommendation, and signs off in 30 minutes is governance.
The discipline:
- Every steering committee paper has at most three decisions, each presented as the recommended option, the alternative considered, and the reason for the recommendation.
- The steering committee meets at fortnightly cadence for engagements under three months, monthly for longer ones. Anything more frequent is operational; anything less frequent is symbolic.
- The engagement manager owns operational decisions. The steering committee owns scope, budget, and direction. The boundary is written down before the first meeting.
Change Control
Scope creep is the silent killer of consulting margins. It rarely arrives as a request to do twice as much work; it arrives as a series of small extensions, each of which seems reasonable in isolation. By week eight you are doing 30% more than the contract called for and writing it off as relationship investment.
The protection is a written change control process from day one. The mechanics:
1. Every change request — from the client or from the consulting team — goes into a change log.
2. Each entry records the change, the originator, the impact on scope, time, and cost, and the decision.
3. Changes above a defined threshold (commonly 5% of fee or one week of effort) require a contract addendum.
4. The change log is reviewed at every steering committee meeting.
Clients respect change control when it is presented as commercial professionalism, not as obstruction. The phrasing matters: "This is a substantive addition that the existing scope cannot absorb. I will scope it as an addendum and bring it to the next steering." This sounds reasonable because it is.
What AI Changes
Generative AI has changed several mechanical parts of consulting project management without changing the underlying discipline.
- Inception drafting compresses from a week to an afternoon. The inception report structure is well-known; AI drafts the standard sections and the engagement manager edits the substance.
- Status reporting compresses from hours to minutes. Meeting notes, weekly updates, and steering papers are generated from raw notes and edited.
- Risk register maintenance becomes a continuous exercise rather than a quarterly one. AI surfaces risk patterns from project artefacts that a human would miss.
- Analysis synthesis does not compress. The hard work of going from data to insight to recommendation is still human judgement, and AI that pretends otherwise produces glossy nonsense.
The new rule: use AI to compress the parts of project management that are pattern-following, and protect the parts that are judgement. The engagement manager who outsources judgement is selling expertise they have stopped exercising.
A Final Discipline: The Weekly Reset
Every Friday, the engagement manager spends 30 minutes alone with the workplan, the risk register, and the change log. The three questions:
1. What slipped this week, and what does that imply for the milestone three weeks out?
2. What risk became more likely or more severe? Does the mitigation still hold?
3. What request from the client this week is a scope change in disguise, and have I logged it?
Run that reset every week and the engagement does not surprise you. Skip it for three weeks and you spend the rest of the project explaining slippages.
Where to Take This Next
- For drafting an inception report end-to-end, see our Document Studio.
- For modelling risk and stakeholder positioning, see Research Frameworks.
- For the contracting side of change control, see the Legal Workshop.
- For a deeper read on a specific phase, the Inception Report guide covers phase one in detail.