Discovery

What happens here

Discovery is the phase that runs from a signed SOW to a validated, defended specification ready for development. The signed contract gives the agency a mandate to investigate; discovery is what produces the artifact that justifies the next phase’s commitments. Skipping or compressing discovery is the leading cause of fixed-price agency engagements blowing through margin.

Five sequential activities make up the phase:

1. Stakeholder Interviews — 1:1 conversations with sponsors, end-users, integrators, and detractors that surface context, constraints, and political realities before the requirements workshop.
2. Requirements Workshops — facilitated sessions where interview themes are converted into concrete, prioritised requirements with explicit scope boundaries.
3. Prototyping & Proof of Concept — time-boxed validation of risky technical, UX, or integration assumptions before they become fixed-price commitments.
4. Estimation & Cost Commitment — the post-discovery re-estimate, reconciled against pre-sales pricing, with a clear commitment path.
5. Discovery Deliverables & Sign-Off — the formal closure ritual that produces the handoff package to Requirements & Design.

The phase typically runs two to six weeks for a mid-sized engagement, though regulated-industry or platform-replacement engagements can stretch to twelve weeks or more. Participants include a delivery lead or principal consultant, a technical lead, a UX lead where design is in scope, and named client stakeholders identified during pre-sales. The outputs — a prioritised feature list and constraint catalogue, a scope statement, prototype findings, and a refined estimate — land in Requirements & Design, where the atomic, testable, traceable FR/NFR document, design artifacts, and architecture decisions are produced from discovery’s validated raw material.

A clean discovery exit is a paid event. The agency invoices on the discovery milestone (typically 15%–30% of total engagement value) regardless of whether the client decides to proceed with delivery. That commercial separation matters: it makes the discovery output a deliverable in its own right, not a pre-sales loss leader.

Best practices

Sequence interviews before workshops. A workshop populated with cold stakeholders defaults to the loudest voice in the room. Conduct individual interviews first — the workshop then becomes the place where pre-surfaced themes are debated and decided, not where requirements are discovered for the first time. Interview-first discovery produces sharper workshops in less calendar time.

Time-box every activity, including prototyping. Discovery has a fixed budget. Stakeholder interviews are 45–60 minutes each. Workshops are half-day or full-day blocks with a published agenda. Prototypes have explicit time-box ceilings (typically 3–10 days) and are abandoned if they exceed it. Open-ended discovery is how a four-week engagement becomes a ten-week engagement and the delivery phase is funded out of margin.

Document decisions in-session, not after. Every workshop produces a decisions log captured live by a designated note-taker, displayed on the screen, and reviewed before the room breaks. Email summaries written the next morning never match what participants thought they agreed to. In-session capture eliminates the “but we said…” conversations that derail delivery.

Treat the re-estimate as a feature, not a problem. Discovery exists in part to produce a defensible estimate. If discovery confirms the pre-sales price, the agency closes with the original number. If discovery reveals that the engagement is materially larger or smaller, the agency triggers the change-control process and re-prices. Mature clients expect this; immature clients need the loop explained at the discovery kickoff. The mechanics live in Estimation & Cost Commitment.

Tip

Distribute the workshop pre-read 48 hours before the session. A pre-read is one or two pages: the workshop’s purpose, the decisions to be made, the materials to review beforehand. Workshops that begin with a 30-minute “context refresh” are signalling that the pre-read was not done — both your fault for not sending it and the participants’ for not reading it. Sending the pre-read on time, and naming who has not opened it, raises the standard of the room.

Caution

Do not start workshops without stakeholder interviews. A workshop with cold participants becomes a discovery session disguised as a decision session — the agency walks out with notes instead of decisions, and the next workshop is needed to re-litigate the same ground. Interviews are not optional; they are the input that makes the workshop productive.

Maintain a parking-lot for out-of-scope items. Workshops surface ideas that do not belong in the current engagement: enhancements, follow-on phases, integration opportunities. Capture them in a parking-lot document rather than letting them either die in the room or silently expand scope. Parking-lot items become candidates for change requests during delivery or for new engagements after launch.

Keep the client visible in the deliverables. Discovery deliverables — the requirements document, the scope statement, the prototype findings, the revised estimate — should be reviewable by the client throughout the phase, not handed over in a final-week document drop. Continuous visibility prevents the closing-day surprise where the client realises the document does not match what they thought was being discussed.

Desired outcomes

By the end of Discovery, the engagement has the following in place:

• A documented, prioritised feature list and constraint catalogue, validated against stakeholder interviews and workshop decisions, ready for formal FR/NFR specification in Requirements & Design — Discovery validates what should exist; R&D produces the atomic, testable, traceable spec that engineers build against
• A signed discovery sign-off document, listing accepted scope, agreed scope exclusions, documented assumptions, and an open-risks log
• A refined cost estimate that either confirms, raises, lowers, or scope-adjusts the pre-sales price — with a defensible methodology behind the number
• A handoff package for Requirements & Design: the prioritised feature list and constraint catalogue, scope statement, prototype findings, decisions log, named stakeholders with roles, and the open-issues backlog
• A commitment path agreed with the client: proceed at confirmed price, proceed at re-priced price, scope-adjust to fit budget, or pause with discovery deliverables retained

What the industry does

Paid discovery vs. included discovery. Mature agencies bill discovery as a separate, paid milestone — either a fixed-price discovery contract that precedes the delivery SOW (the discovery-first model) or a discovery phase line-item inside a phased SOW. The deliverables are owned by the client regardless of whether they proceed with delivery. Less mature agencies bundle discovery into the delivery price as a free consultation, anchored by the original pre-sales number. Paid discovery is the more defensible model: it produces a specification that justifies the delivery price, aligns expectations, and gives both parties an exit ramp. Free discovery consistently anchors delivery to under-quoted pre-sales numbers because the agency cannot credibly re-price after work has already been done.

Time-boxed vs. open-ended discovery. Time-boxed agencies set a fixed calendar duration for discovery (two weeks, four weeks, six weeks) regardless of how much remains “to learn” at the deadline. The deadline forces prioritisation; open questions become entries on the open-risks log carried into delivery. Open-ended discovery continues until the agency feels confident enough to estimate — which in practice means it continues until budget exhaustion. Time-boxed discovery is the dominant model in mature agencies; open-ended discovery survives mainly in research-led engagements where ambiguity is the point.

Heavyweight requirements documents vs. lightweight specs. Some agencies produce 80-page formal requirements documents — full FR/NFR catalogues, ISO/IEEE-style structure, traceability matrices. Others produce 15-page lightweight specs — narrative scope statement, prioritised feature list, accepted assumptions, open risks. Heavyweight is appropriate for regulated industries (healthcare, financial services, government) where the document is itself a compliance artifact. Lightweight is appropriate for most commercial agency work, where a long document slows delivery without changing what gets built. The format choice is a function of the client’s industry, not the agency’s preference.

Discovery as gateway vs. discovery as continuous activity. Traditional agency lifecycles treat discovery as a discrete phase that closes before development begins. Continuous-discovery shops keep discovery activities running in parallel with delivery — ongoing user research, iterative requirements refinement, evolving prototypes. Continuous discovery suits product-development engagements where the client owns long-term roadmap and the agency is one of multiple workstreams. Phase-gate discovery suits fixed-scope project work where the contractual commitment requires a stable spec. Most agencies pick one and apply it consistently; agencies that mix the two confuse both their delivery teams and their clients.