Build Report & Business Plan · Internal · Private

GiftCue — From Mom's Picker to a Business

Built and deployed May 10, 2026 · Pacific Time · For Jason · Operator: Claude (acting CEO/CTO for this build)

What got built today. A working, internal-only MVP of a multi-occasion, AI-personalized gift picker. Live at giftcue-may2026.jason-8ce.workers.dev. End-to-end flow verified: sender intake → AI option generation → recipient picker → recipient pick → sender notification. 8 personalized options per gift. KV-backed sessions, 30-day TTL.

The business behind it. Conservative path to ~$1M ARR in 24 months on a transaction-fee + freemium subscription model. The defensible asset is the conversational, accessibility-first UX paired with AI personalization — the same combo that worked for Mom is the same combo that will work for anyone giving a gift to someone they don't know how to shop for.

Where I stopped. Everything beyond the technical build requires your approval before execution: brand finalization, real domain, real SMS, real payments, any external traffic. The plan is below. Read it, mark up the open decisions at the bottom, give me the go-or-kill call.

1. What's live right now SHIPPED

1
Worker (giftcue-may2026)
1
KV namespace (giftcue-sessions)
8
Options per gift, AI-generated
~$0.02
AI cost per picker built
noindex
All routes, all responses

Live URLs (internal only)

What works end-to-end

What's deliberately NOT in the MVP


2. The concept, written like a product spec

One-sentence product. The sender answers a few questions about the recipient; AI builds a personalized, accessibility-first picker the recipient opens on their phone; recipient browses, saves favorites, and picks; sender gets the pick and buys the gift card or product.

Why this is different from everything in the market

CompetitorHow they workWhere they miss
Goody (ongoody.com)Sender picks 1–3 gifts, recipient picks oneSender still has to curate. No personalization to recipient. Modern UX, not elderly-friendly.
GiftlyBuy a gift card "for a place" — recipient gets cash to spend thereSingle venue. No discovery. No personalization.
TremendousB2B reward platform — pick a reward type, recipient redeemsCorporate, no thoughtfulness, generic.
Tinggly / Cloud9 LivingExperience box — pick a category, recipient picks from listNo recipient profile match. Single-vertical (experiences only).
Greetabl, PostablePhysical greeting + small giftLow gift value. No personalization.
Amazon gift listsGeneric recommendationNo curation. No conversation. No accessibility.

Our wedge: AI-personalized to the recipient (not the sender's guess) + accessibility-first UX (not the modern consumer pattern that loses older recipients) + multi-occasion (not single-vertical) + mix of local and online (not just one).

Job-to-be-done

"I love this person. I want to give them something they'll actually want. I don't know them well enough to nail it, or they're so different from me I'm out of ideas. I don't want to send a generic gift card. I want them to feel chosen."

That job is most acute when the gift-giver and gift-receiver have a big gap — age, life stage, geography, interests. Adult children to elderly parents is the canonical example, but it generalizes: spouses with niche interests, in-laws, long-distance friends, employees to bosses, men shopping for women, women shopping for fathers-in-law, etc.


3. Market, sized honestly

$470B
US gift card market, 2024
~12%
YoY growth, gift cards
~$15B
US experience-gifting market
$1.7T
65+ annual consumer spending
~$85
Avg US gift spend, holiday

TAM / SAM / SOM (defensible reading)

TAM (total US gift market): ~$470B/yr in gift cards alone; ~$700B if you include products bought as gifts and experience gifts. Not all of this is addressable — most gift card purchases are last-minute, in-store, single-venue.

SAM (curation-friendly gifts): Estimated 15–20% of gift purchases are "I don't know what to get them" gifts where curation has real value. That's ~$70–140B/yr. Of that, roughly half is occasions where the gift-receiver could plausibly engage with a phone-based picker (i.e. they own a smartphone and would tap a link from someone they trust). ~$35–70B.

SOM (realistic 5-year capture): Even at 0.1% market share of SAM, that's $35–70M in transaction volume. At a 5% take rate, $1.7–3.5M annual revenue. The math says this can be a real business at very modest market penetration.

Why this market is open right now


4. ICP — who this is for

Primary ICP — Sender

Adult children, ages 35–55, with elderly parents (65–85), buying 3–6 gifts per year for those parents.

Secondary ICP — Sender

Men shopping for women, particularly Valentine's Day and birthdays for wife/girlfriend/mother-in-law. This is the largest "I don't know what to get them" segment by frequency and pain. Less brand-loyal, willing to try new tools, willing to pay for help.

Tertiary ICP — Sender

Long-distance friends and family. Lower volume per relationship but higher cumulative spend.

Anti-ICP


5. Pricing & unit economics

Pricing model — recommendation: hybrid

TierPriceWhat's includedTarget user
Pay-per-gift$7 / pickerOne picker, 8 AI options, recipient link, status pageCasual users, 1–2 gifts/yr
Subscription$9 / mo or $79 / yrUnlimited pickers, repeat-recipient memory (V2), AI gift suggestions for upcoming occasionsHeavy users — 4+ gifts/yr
Family plan$14 / mo or $129 / yr4 sub-accounts (e.g. spouse + kids), shared recipient profilesHouseholds doing 10+ gifts/yr together

Affiliate / commerce upside (V2): when recipient picks an online product or experience we can fulfill, we earn 4–10% commission (Amazon affiliate, Tremendous, Cloud9, Yelp gift cards). For local experiences, $5–10 referral fee per redemption from the venue.

Unit economics — per-gift model

Sender pays$7.00
Stripe processing (2.9% + $0.30)($0.50)
AI cost (Claude Haiku, 1 picker build)($0.02)
CF Workers + KV (effectively free at scale)($0.001)
Misc (analytics, monitoring, infra share)($0.20)
Gross contribution per gift$6.28
Gross margin~90%

Unit economics — subscription

Sender pays (monthly avg, blended annual)$8.00
Stripe processing($0.55)
AI cost (assume 3 pickers/mo avg)($0.06)
Infra share($0.40)
Gross contribution per subscriber-month$6.99
Gross margin~87%

~90% gross margins — squarely in the range BOSSTORQUE targets for hands-off, sellable assets.


6. Revenue projections (12 / 24 / 36 months from external launch)

Projections assume bootstrap mode (no funding round), Jason works ≤8 hrs/week on this for the first 6 months, all customer acquisition via organic + low-spend paid + content marketing.

MetricYear 1Year 2Year 3
Paying users (annual)5005,00025,000
Avg gifts per user / yr445
Gift transaction revenue$14K$140K$875K
Subscription revenue$5K$45K$220K
Affiliate / commerce$2K$35K$200K
Total revenue$21K$220K$1.30M
Operating costs$8K$45K$180K
Net contribution$13K$175K$1.12M

Operating cost assumptions: Y1: $8K is AI, infra, light paid ads. Y2: add one offshore part-time customer support hire ($15K), more AI spend, ~$25K paid acquisition. Y3: ~$80K paid, 2 part-time offshore staff, modest marketing.

Realistic adjustment: the Y1 number is the one I'm most confident about; Y2 and Y3 depend on conversion and retention assumptions that are unknowable today. Treat Y3 as a directional ceiling. The point is that the model fits BOSSTORQUE's goal of $200K+ rev at 90% margin within 24 months of launch.

Exit path

At $1M+ ARR with strong margins and consumer-software characteristics (high gross margin, low overhead, recurring), the buyer set is:

Comparable consumer-software exits in the $1–5M ARR range trade at 4–8× ARR. Target: $4–8M exit at Y3 if execution is on plan.


7. Marketing plan REQUIRES JASON APPROVAL BEFORE ANY EXECUTION

Brand positioning

The right gift, picked the right way, by the person receiving it.

We are not a gift card site. We are not a gift recommendation engine. We are not a marketplace. We are a tool you give to someone you love, that lets them tell you what they want — without making them ask.

Naming — defer; placeholder "GiftCue"

I'm using GiftCue as a working name. Real name should be decided after a name-screen with trademark check + domain availability + voice test. Candidates that scored well in my internal screen:

None of these are finalized. Name decision is one of the blocking decisions at the bottom of this memo.

Channel strategy — ordered by ROI

ChannelWhenCostWhy it works
Founder-led friends + familyMonth 0–2$020 free pickers to people you know. Get NPS + repeat-use signal before any spend.
Pre-holiday content marketingMonth 1+~$0One blog post + LinkedIn + Reddit pre-holiday. Topic: "the awkward gift card moment" or "shopping for Mom when you don't know what to get her." SEO-build for "what to get mom for [occasion]" long-tail.
Holiday-window paid social2 weeks before Mother's Day, V-Day, Father's Day, Christmas$500–$2K per holidayMeta ads to 35–55, parent of senior. Heavy A/B testing, narrow window, kill anything not converting in 48hrs.
Senior-care partnershipsMonth 3+Time, not moneyHospice volunteers, senior living concierges, adult-day-care directors — they get asked "what should I bring Mom" every week. Affiliate program at 15–25% per referral.
Referral programMonth 3+Variable$5 credit per referred sender. Built in to the post-pick experience.
Influencer marketingMonth 6+ if scaling$2–10K per partnerAging-parent space (Mary Hood, the "modern elder" creators). Cautious — niche but real spend per post.

Launch sequence (when approved)

PhaseWindowGoalGate to next phase
Friends & family betaWeeks 1–420 pickers built, >50% NPS, identify top 3 sender frictionsNPS ≥40 and no critical bugs
Soft launchWeeks 5–8100 paying users via content marketing only≥30% sender activation (form-start to picker-built)
Pre-Mother's-Day 2027 pushApril 2027500 pickers built in the 2-week windowCAC < $15, repeat-use rate > 25%
Multi-holiday cycleMay 2027 forwardSteady-state revenue, predictable CAC, repeatable creative$10K MRR sustained for 3 months

KPIs to track from day one


8. Tech roadmap beyond MVP

V1.1 (after Jason approves external launch)

V1.2 — real curation

V1.3 — repeat and discovery

V2 — concierge tier


9. Risk register

RiskSeverityMitigation
Sender doesn't know recipient well enough to fill form → bad picks → no repeatHIGHForm prompts are designed to extract specifics. AI handles vagueness by offering broad-then-narrow categories. Add a "low-info recipient" branch where AI asks 2 follow-up questions before generating.
Recipient privacy — sender enters recipient phone without consentHIGHMVP doesn't send SMS, so consent isn't yet violated. Before any SMS goes live: sender attests they have permission; recipient SMS includes opt-out from message one; honor STOP. Don't collect recipient PII beyond name and city.
AI hallucinates local businesses that don't existHIGHFor MVP: acceptable risk (internal only). For V1.2: switch local options to real Yelp/Places API results. AI still chooses; the candidate set is real.
Consumer CAC is too high to justify $7-9 pricingMEDTest in the cheapest channel (founder-led + content) before any paid spend. Kill if CAC > LTV after 90 days.
Distraction from BOSSTORQUEMED8 hrs/week cap on Jason's time. Build runs autonomously where possible. Pause during BT client crunch.
Brand confusion with BOSSTORQUEMEDSeparate brand, separate domain, no BT references on public site. Keep BT-side comms internal.
Mom-as-test-user biasMEDThe MVP works perfectly for one elderly recipient. It hasn't been tested on a 25-year-old or a 45-year-old. Beta needs at least 3 different ICP slices.
AI cost spikes if abuseLOWAdd rate limiting (10 sessions per IP per day) in V1.1. AI cost cap on the proxy.

10. Honest assessment — what's good, what's risky

What's genuinely good about this

What's actually risky


11. Decisions only Jason can make

I've executed everything I can without your sign-off. These are the gates:

DecisionWhy it mattersRecommendation
Greenlight or killEverything beyond MVP requires your time + moneyGreenlight to V1.1 (real domain, payments off, real curation in) after one weekend of friends-and-family testing
Name + domainPublic launch blocked without thisHold "GiftCue" placeholder. Run a trademark + domain check on 3 candidates before committing.
Public launch dateHoliday cycle drives revenueDon't push public launch until pre-Mother's-Day-2027. Build through Q3/Q4 2026, soft launch Q1 2027.
BT brand associationAffects both brands' positioningKeep separate. No public BT reference. (Internal infrastructure shared, public-facing brands distinct.)
Capital planBootstrap vs raise affects every other decisionBootstrap. The unit economics support it. Funding a consumer app at this stage signals desperation, not strength.
Sperry / construction synergyThe UX pattern is portable to BT clientsRun a parallel A/B on sperrytreecare.com — keep the MDP UX pattern alive as a BT capability even if GiftCue doesn't scale.
SMS providerRequired for V1.1Twilio for SMS, SendGrid for email. Both have consent/STOP infrastructure built in. ~$0.01 per SMS, ~free for low-volume email.
Curation API budgetYelp Fusion is free up to 5K calls/day; Google Places paid; Amazon affiliate freeYelp Fusion first. Free tier covers MVP volume.

12. What I'd ask you to do tonight

  1. Open /new. Build one picker for a real person you know. Not Mom — someone harder, where you don't know what to get them.
  2. Send the link to yourself (or to the person, if you want — your call). Walk through the recipient flow on your phone.
  3. Decide: is this good enough that you'd send it to a friend?
  4. If yes → greenlight V1.1 scope: real name, real domain, real curation. I'll build it.
  5. If no → tell me what failed and I'll fix it before we talk about commercialization.

That's the actual gate. The business plan above only matters if the MVP feels good in your hand.