Write client proposal

Write Client Proposal

Turn an understood client problem into a written offer the client can say yes to — priced on value, scoped in tiers, with a timeline that survives the feasibility check and a change process that protects the work.

When to use / when NOT to use

Use when:

• A client or prospect asks for a price, quote, proposal, or statement of work
• Scoping any paid engagement — new client, new phase, or expansion of existing work
• The client's ask is vague ("we need a new website") and you must respond with something priced — this skill's Step 0 decides between a full proposal and a paid discovery proposal
• Reviewing a drafted proposal before it goes to the client

Do NOT use when:

• Deciding whether the client's problem is worth solving at all, or sizing it → identify-business-problems (this skill consumes its sizing and feasibility outputs)
• The client hasn't articulated goals yet and you have access to them → conduct-stakeholder-interviews first; proposals written without discovery are guesses with a price tag
• Writing the differentiation story ("why me over others") → position-product supplies it; this skill only places it
• Sending a routine status update to an existing client → that's an email, not a proposal (BLUF rules still apply via writing-voice)
• Pricing a product (subscriptions, paywalls) rather than your own services → design-growth-surfaces

Pipeline position: conduct-stakeholder-interviews → identify-business-problems (size + feasibility) → write-client-proposal → signed scope → work → tell-case-study-story (afterward, with permission).

The method

Step 0 — Qualify the bid (before writing a word)

Not every request deserves a proposal. Place the engagement on the quadrant first:

                LOW REWARD                HIGH REWARD
LOW RISK    →   Small fixed quote, fast   Full tiered proposal — protect time for it
HIGH RISK   →   Decline or refer          Paid discovery first, then re-propose

• Reward = fee + relationship/portfolio value, honestly weighed. A tiny fee with a genuinely reusable case study can still be HIGH; "exposure" alone is LOW.
• Risk is HIGH when ANY of: the problem is unsized (nobody knows what it's worth to the client), scope is undiscoverable from the outside, the client wants a price before defining the problem, payment terms are resisted, or delivery would collide with existing committed capacity.
• Unsized problem + client wants a number anyway → propose a small fixed-price discovery engagement (interviews, data pull, sized findings, written recommendation). Discovery is real deliverable work, never free pre-sales labor. Its output feeds identify-business-problems, and the follow-on proposal cites its findings.
• Capacity fails (see Step 3) → decline, refer, or offer a later start date. Never bid on time that doesn't exist.

Step 1 — Anchor the value before the scope

The proposal sells an outcome, not deliverables. Before listing anything you'd build:

1. Name the client's milestone — outcome + number + date, in their words from discovery ("fully booked chairs by the January location opening", not "grow the business"). No milestone surfaced → get one (conduct-stakeholder-interviews) or propose discovery. A proposal anchored to nothing is decoration.
2. State the value at stake — the bottom-up sizing from identify-business-problems (count × share × value, as a range, sources named). This number is the pricing anchor in Step 3 and the opening line of the proposal.
3. Run every external claim you plan to make through craft-critique's evidence-discipline protocol (it owns the cite/get/flag disposition — apply it, don't restate it). A proposal is external-facing; under-evidenced claims either get evidenced before sending or get cut.

Step 2 — Build three scope tiers

One take-it-or-leave-it scope turns the conversation into yes/no on price. Three tiers turn it into which one — and let the client self-select on budget without negotiating your rate.

Rules for tiers:

• Each tier is a complete outcome, not a crippled version of the one above. The smallest tier must still move the client's milestone on its own. If it can't, it isn't a tier — cut it.
• Name tiers by outcome, never Bronze/Silver/Gold ("Booking flow, live" / "Booking flow + intake automation" / "Full front-door rebuild").
• Differentiate on scope, not quality. Every tier gets the same craft. What varies: how much of the problem gets solved, how fast, and how much ongoing support follows.
• Three is the number. Two reads as cheap-vs-expensive; four+ forces the client to do your scoping work.
• Recommend exactly one — decision-support shape: options presented, one recommended, with specific reasoning tied to their milestone. A proposal that presents tiers without a recommendation makes the client do the analysis they're paying for.

Step 3 — Price each tier

Run this decision tree:

Is the scope defined well enough that YOU control execution?
├── YES → FIXED PRICE per tier (default — you keep efficiency gains)
└── NO  → Is discovery the missing piece?
          ├── YES → small fixed-price DISCOVERY engagement first (Step 0)
          └── NO (genuinely open-ended, e.g. ongoing advisory)
                → weekly/monthly RETAINER with a defined capacity
                  (hourly billing is the last resort — it prices
                   your time, caps your upside, and invites
                   line-item auditing of your invoices)

Then set each fixed price between a floor and an anchor:

1. Floor = realistic hours (Step 4's estimate, buffer included) × the minimum rate that makes the work worth taking over other uses of the same hours (other paid work, portfolio time — the opportunity cost is real even when unbilled). Never price below floor. A prestige project priced below floor is a Spend-without-ROI decision — make it consciously and record it (write-decision-rationale), don't drift into it.
2. Anchor = the value at stake from Step 1. A fix that recovers a recurring monthly leak comfortably supports a price of one to a few months of that leak; a deadline-shaped problem (their launch, their lease decision) supports more. If the defensible price barely clears the floor, the problem is too small for the engagement — shrink the scope or decline.
3. State one number per tier. Ranges in a proposal read as "the real price is the bottom of the range." Do range-thinking privately; print a number.
4. Discount = remove scope, never cut the price of the same scope. "Same work, less money" reprices all future work in one sentence. If budget is the blocker, the smaller tier IS the discount.
5. Third-party costs (hosting, API usage, fonts, stock, plugins) pass through at cost, named in Assumptions — never silently absorbed into the fee.

Step 4 — Feasibility-check the timeline (before promising anything)

A promised date that fails the math is a client-drop risk waiting to fire. Run the feasibility check from identify-business-problems (Step 7) on every tier — estimate realistically from actuals, ×1.5 for first-of-kind work, count real weekly capacity after standing commitments, and confirm:

estimate ≤ available hrs/week × weeks × 0.8   (minimum 20% buffer, always)

• Fails? The proposal changes — longer timeline, later start, or smaller tier. Sending it anyway is the failure.
• Name what it displaces. Capacity is zero-sum; a signed proposal that only adds work is a lie you told yourself first.
• Quote calendar dates from client-dependency zero — the clock starts when their inputs arrive (content, access, approvals), not when the contract signs. Say so in the proposal.
• Freeze the estimate at send time; log actuals at the end. The gap is your calibration data for the next proposal — never revise mid-project to look on-plan.

Step 5 — Assumptions, exclusions, and the revision allowance

Every dispute in a fixed-price engagement traces back to something one side assumed and nobody wrote down. Write down:

• Assumptions — what must be true for the price and dates to hold: client delivers content/access/credentials by named dates; feedback returned within N business days (late feedback slips the timeline day-for-day — say this sentence verbatim); one named decision-maker consolidates feedback; third-party services keep working as documented.
• Exclusions — the adjacent work a client will predictably assume is included: content writing, logo/brand design, ongoing maintenance, SEO, staff training, other pages/flows "while you're in there." Name the nearest ones explicitly.
• Revision allowance — a number, per tier (e.g., two consolidated revision rounds per deliverable). "Revisions until you're happy" is unlimited scope wearing a friendly face.

Step 6 — Change control and payment terms

Change control (put this in every proposal, run it without apology):

Client asks for anything outside the written scope
  → acknowledge same day, log it
  → send a written change order: what it adds, price delta, timeline delta
  → nothing outside scope gets built until the change order is approved

Absorbing "one small thing" silently doesn't buy goodwill — it teaches the client that scope is soft. The change order is a 10-minute document, not a confrontation.

Payment terms:

• Deposit before any work starts — 50/50 for short engagements, thirds (start / midpoint / delivery) for longer ones. Never 100% on completion — that makes you the client's lender.
• Final deliverables (handoff, credentials, source) transfer on final payment.
• A pause/kill clause: if the client pauses beyond N weeks or cancels, work completed is billed at the tier's effective rate and the restart gets rescheduled around then-current capacity — a pause is not a free reservation.
• Proposal expiry: price and dates valid for 14 days. Capacity moves; an un-expired proposal is an option contract you gave away free.

Step 7 — Write it (BLUF, two pages, their words)

Load writing-voice (Professional register) and write in this order — the recommendation and price appear on page one, never buried:

# Proposal — [outcome, in the client's words] · [date] · valid until [date]

## The outcome                       ← BLUF: what changes for their business,
[2–3 sentences: their milestone,       anchored to the milestone + value at stake]
what this engagement moves, and the recommended option + its price]

## The problem as we discussed it    ← their words from discovery, played back;
[3–5 sentences + the sizing]           this is where they decide you listened

## Options
### [Tier 1 name — outcome] — $X · [duration]
[What's included — outcomes first, deliverables under them]
### [Tier 2 name] — $Y · [duration]   ★ Recommended — [one sentence why,
### [Tier 3 name] — $Z · [duration]      tied to THEIR milestone]

## Timeline
[Phases with calendar dates, clock starting from client-dependency zero]

## Investment & payment
[Chosen tier's price · payment schedule · pass-through costs]

## Assumptions & what's not included
[Step 5 output — bulleted, unambiguous]

## Changes
[Step 6 change-control paragraph]

## Relevant work
[1–2 shipped, checkable proofs relevant to THIS problem — pull current
numbers at send time, never reuse figures from an old proposal;
positioning language comes from position-product]

## Next step
[One action: reply picking an option / 20-min call — with a date]

• Under ~2 pages for typical freelance scope. The proposal's length signals the project's weight — a 12-page proposal for a 3-week build reads as insecurity.
• Their words beat your taxonomy: if discovery said "patients give up on the phone," the proposal says that — not "suboptimal conversion funnel."

Step 8 — Gate it, then record it

1. Run craft-critique's pixel-polish ship gate — a proposal is external-facing work; typos and misaligned tables price you down before the number is read. Every metric in "Relevant work" passes the evidence protocol.
2. Read it once as the client — a busy owner giving it 90 seconds. Does page one alone let them say yes?
3. Record the pricing/scope decision via write-decision-rationale, including What Would Change This (e.g., "priced assuming 10 hrs/wk freelance capacity — re-quote if your available capacity changes").
4. Proposal is Tier 3 in ARIA terms: Dinesh reviews and sends it himself. Drafts never go out unreviewed.

Proposal lint (every proposal, before send)

• Milestone named in the client's words (outcome + number + date) in the first section
• Value at stake stated, bottom-up, sourced or flagged (from identify-business-problems)
• Exactly three tiers; each a complete outcome; exactly one recommended, with milestone-tied reasoning
• One number per tier — no ranges, no "TBD", no "starting at"
• Every tier's price ≥ floor; anchor logic checked against value at stake
• Feasibility math run and passed for the recommended tier; displacement named (privately)
• Timeline in calendar dates, clock pinned to client-dependency zero
• Assumptions, exclusions, and a numeric revision allowance present
• Change-control process and payment schedule (deposit ≥ first third) present
• Pause/kill clause + 14-day expiry present
• Proof claims current and checkable — no stale numbers carried over
• ≤ ~2 pages; BLUF holds — option + price on page one
• craft-critique ship gate passed; decision recorded; Dinesh sends it himself

Worked example (fictional — a two-location physiotherapy clinic)

The ask: "We need a new website." Step 0: problem unsized → HIGH risk. Owner is reachable → ran discovery instead of quoting. Discovery found: new-patient intake happens by phone; front desk misses calls during sessions; the owner's real milestone is "both locations at full schedule before the March associate hire." Sizing (via identify-business-problems, from their call log): ~30 missed intake calls/month × ~40% never call back × ~$700 first-treatment-plan value ≈ $6–10k/month leaking (range; call-back share under-evidenced, flagged).

Tiers built:

Tier	Outcome	Price	Duration
Online intake, live	Booking + intake form replacing phone-only intake, both locations	$4,800	3 wks
★ Intake + follow-up automation	Above + automated reminders and missed-call text-back	$7,900	5 wks
Full front-door rebuild	Above + site rebuild + on-site conversion pass	$14,500	9 wks

Pricing logic (middle tier): realistic estimate 38 hrs; missed-call text-back is first-of-kind for this stack → ×1.5 → 57 hrs. Floor at a $90/hr opportunity-cost rate ≈ $5,100. Anchor: recovering even the low bound ($6k/mo) pays the fee back inside month two — $7,900 sits defensibly between floor and anchor. Feasibility: 10 freelance hrs/wk × 8 wks × 0.8 = 64 usable hrs ≥ 57 → passes; quoted 5 calendar weeks of build across an 8-week window, clock starting when the clinic delivers scheduling-system access. Displaces: the portfolio's next case study slips two weeks — accepted consciously, recorded.

Recommended tier 2 in the proposal: "Recovering missed intake calls is what moves you toward full schedules by March; the site rebuild can follow once the leak is closed." Assumptions included: clinic provides scheduling-system access by day 1, consolidated feedback within 3 business days (late feedback slips day-for-day), two revision rounds per deliverable; SMS costs pass through at cost. Excluded by name: content writing, brand refresh, ongoing maintenance (offered separately as a retainer after launch). Payment: 50% to start, 50% at delivery; 14-day expiry. What would change this (recorded): if their call log shows the call-back share is actually >80%, the leak shrinks below the anchor — re-scope before signing.

Anti-patterns / red flags

Red flag	Why it fails	Do instead
Quoting a price before the problem is sized	You're pricing blind; the number is a guess with your name on it	Step 0: paid discovery first, or size it via identify-business-problems
Hours × rate as the price ceiling	Prices your time, not the outcome; punishes your own efficiency	Floor from hours, price from value at stake
One take-it-or-leave-it scope	Turns the decision into yes/no on price	Three complete-outcome tiers, one recommended
Tiers without a recommendation	Client does the analysis they're paying you for	Recommend one, reasoning tied to their milestone
A price range or "starting at…"	Client hears the bottom number, budgets the bottom number	One number per tier
Optimistic timeline to win the bid	Deadline miss + client drop later costs more than losing the bid	Feasibility math first; the date survives or the proposal changes
Timeline clock starts at signature	Client delays become your lateness	Pin dates to client-dependency zero, in writing
"Unlimited revisions" / no revision number	Unbounded scope wearing a friendly face	Numeric revision allowance per tier
Absorbing "one small thing" off-scope	Teaches the client scope is soft; creep compounds	Same-day change order, every time
Discounting the same scope	Reprices all your future work in one sentence	The smaller tier IS the discount
100% payment on completion	You become the client's lender	Deposit up front; deliverables transfer on final payment
No expiry date	A free option on your future capacity	Valid 14 days
Proposal written in your taxonomy	Client can't see themselves in it	Their discovery words, played back
Reusing last year's proof metrics	Stale numbers fail the evidence protocol — and can fail in the room	Pull current numbers at send time, every time
12 pages for a 3-week build	Length reads as insecurity, not rigor	~2 pages, BLUF, page-one decision

Boundaries

• identify-business-problems owns opportunity sizing, the risk-reward quadrant mechanics, and the feasibility/capacity check (its Step 7) — this skill applies them to bids and timelines, citing that skill rather than restating the full procedure.
• conduct-stakeholder-interviews owns discovery — the milestone and the client's words come from there.
• position-product owns the differentiation story placed in "Relevant work"; tell-case-study-story owns the long-form proof used after the work ships.
• writing-voice owns register and prose (Professional register for proposals); craft-critique owns the evidence-discipline protocol and the pre-send ship gate — reference both, never restate.
• write-decision-rationale owns the decision-record format (including What Would Change This) for pricing and scope calls.
• design-growth-surfaces owns pricing of products/offers inside a client's business; this skill prices only Dinesh's own services.

Sources

• ARIA operating DNA — not a fork, not NN/g-grounded. The method encodes frameworks Dinesh runs by, so its authority is his own operating standards, not external citations: the Risk-Reward quadrant (Step 0), Milestone Anchoring — nothing floats (Step 1), value-anchored pricing and the decision-support tier shape — options presented, exactly one recommended (Steps 2–3), and the feasibility/capacity check with the mandatory ≥20% buffer (Step 4).
• Borrowed protocols keep their owners and are referenced, never restated: opportunity sizing + the quadrant mechanics + the feasibility/capacity math are identify-business-problems' (its Step 7); discovery, the milestone, and the client's own words are conduct-stakeholder-interviews'; the evidence-discipline protocol (cite / get / flag) and the pre-send pixel-polish ship gate are craft-critique's; the differentiation story is position-product's and the long-form proof is tell-case-study-story's; the decision-record format with its mandatory What Would Change This is write-decision-rationale's; register and prose are writing-voice's (Professional).
• The worked example is fictional — the two-location physiotherapy clinic and every figure in it are illustrative, not a real Dinesh client or engagement.