Dogfooding: using our own product to onboard creators

Dogfooding means a company uses its own product, for real, before and while its customers use it. Distribution Club is its own first brand, its own first creators and its own first ops team. This page explains what that means week by week, what it costs, how it doubles as the way we recruit creators, what it must not distort, and how we measure it.

Draft v0.1 · Last updated 4 September 2026

On this page

1. 1. Why we dogfood
2. 2. Brand zero
3. 3. Preset G, "Creator story"
4. 4. The founder as creator, brand and ops
5. 5. First-gig guarantee
6. 6. Payout day as a monthly event
7. 7. Referral loop
8. 8. Friction capture
9. 9. Budget and cost per creator
10. 10. Guardrails
11. 11. Metrics and targets
12. 12. Week by week plan
13. 13. Risks of dogfooding and how we handle them

1. Why we dogfood

A few words first. A brand is the paying customer: a company that wants short videos made about its product. A creator is the person who makes and posts the video. A campaign is what a brand creates: a brief, a 10-item checklist, a fund and some settings. A gig is one unit of work for one creator inside a campaign: one video on one account. An offer is the notification that sends a gig to a creator, who can accept or decline. Ops is the platform's own operations team, which reviews videos, watches for fraud and answers support. A dogfood campaign is a campaign funded by Distribution Club itself, run through the normal product, used to test the product and to recruit creators. Every dogfood campaign promotes a software product, the founder's own tools or a friendly indie developer's product, never a food, restaurant or local business (brief section 50 item 13); this matches V1's customers, indie developers, indie hackers and small SaaS companies. Preset G's "Creator story" (section 3) is the recruiting format used every week and is always about Distribution Club itself. There are three reasons we do this.

To find friction before brands do. Friction is anything that slows a person down or makes them unsure: an offer screen that does not show the deadline, an upload that fails near the end, a review note that says "changes requested" without saying what to change. A bug found by the founder on a test phone is cheap. The same bug found by a pilot brand whose $5,000 campaign is stuck is expensive, because it costs the brand's trust as well as the fix. So the founder and 20 friendly creators use the product on real accounts with real money from day 13 of the MVP sprint, 19 September 2026, and every piece of friction is logged the same day it is hit.

To make creator acquisition a campaign like any other. Most marketplaces recruit creators with ads or with outreach. We recruit creators by paying creators to tell other creators about the product, through the product. A creator gets an offer, makes a short honest video about their own experience, uploads it, the platform reviews it and the creator posts it on their own account for the platform to verify, and the creator is paid by the same CPM rate card, the same floor and cap, and the same payout cycle as a gig for any other brand. The caption carries the creator's referral link. The recruiting machine and the dispatch machine are the same machine, so every dollar spent on recruiting also exercises the product. This matters more than usual because the beta opens worldwide on 25 January 2027, not in one country first: a dogfood campaign and a referral link cost the platform the same expected earnings whether the creator is in Manila or Miami, so dogfooding and referral, not paid ads, are the main tool for building a global creator pool.

To prove the product pays. A creator who signs up does not yet believe the product is real. Belief arrives when money arrives. The wallet is the creator's balance inside the app. A CPM is the rate a creator earns per 1,000 real views their post gets, set by their league; 20% of expected earnings, the floor, is released into the wallet the moment the post is confirmed live. A payout is money moved from the wallet to the creator's bank account on the 1st of the month. A creator's first payout is the moment they decide this is a real way to earn. The dogfood programme is built to get every new creator to a completed gig, a wallet balance and a first payout as quickly as possible, and then to turn that payout into another video.

2. Brand zero

Distribution Club has a brand account like any other brand. We call it brand zero. It gets no special treatment: the same brand dashboard, the same fund (the money deposited for a campaign, held by the licensed payments partner and never in the platform's own operating account), the same platform review against the 10-item checklist, the same "#ad" disclosure and paid-partnership label added automatically at posting, the same CPM (set by audience geography, platform, category and league) with a floor released at publish and a maximum payout per post, and the same automatic payout on the 1st of every month. If brand zero cannot do something in the dashboard, no brand can, and that is the point.

When it starts

Day 13 of the MVP sprint, Saturday 19 September 2026, not week 13 of the old fourteen-week Phase 1 build (brief section 28). Brand zero funds its first dogfood campaign that day, on the founder's own account and the 20 friendly creators' accounts, before the first pilot brand campaign is even funded (day 14, 20 September 2026). Brand zero then funds one dogfood campaign every week through the end of Phase 1 build-out, the whole of Phase 2 (26 March 2027) and the end of Phase 3 (25 June 2027). Phase dates are on the Rollout plan page.

Cadence

The very first campaign, on day 13, is funded on a Saturday because that is where day 13 of the sprint falls; every week after that it is funded on the Monday so dispatch runs across the whole week (assumption), starting Monday 21 September 2026. One dogfood campaign a week, using Preset G "Creator story" (section 3). Offers go out in waves and expire 30 minutes after they are sent, exactly as described on the How it works page.

Budget

$2,000 a week from day 13 of the MVP sprint (19 September 2026) through the whole of Phase 1 build-out and all of Phase 2. $5,000 a week during Phase 3. These are deposits, the same as any brand pays, worked out from what one gig takes from the fund; 20% of each deposit becomes the platform's revenue at once, and the other 80% is the creator pool the gigs draw from. Unspent fund stays as fund balance and rolls into the next week's campaign. The budget is a ceiling, not a spending target.

Who runs it

The founder creates and funds the campaign in the brand dashboard with a real card. At MVP the founder reviews every dogfood submission personally in the ops console, the same review the founder gives every other submission on the platform (brief section 29); the founder's own review of these gigs is part of the dogfood loop, since it means the founder has been through offer, upload, review and payout before asking anyone else to trust the review process. The founder keeps reviewing until automated checklist scoring is trusted at volume; an ops reviewer takes over routine review only once the hire trigger in Rollout plan, section 10 is crossed. The founder runs the monthly payout batch from the console on the 1st. Nobody uses an admin shortcut or a script for any of these steps.

What $2,000 buys. Take a Bronze gig with a typical Tier 1 audience. V1's ten categories all carry category factor 1.6, 60% higher than the old food-example rate for the same views, so average expected earnings across Bronze 1 and Bronze 2 come to $50.00. What a gig takes from the pool is expected earnings plus the $2 running cost: $50.00 plus $2 is $52.00. The weekly pool is 80% of the deposit: $2,000 times 0.8 is $1,600. $1,600 divided by $52.00 is about 31, so the weekly pool buys about 31 Bronze gigs. In Phase 3, at $5,000 a week, the pool is $4,000; $4,000 divided by $52.00 is about 77, so the pool buys about 77 Bronze gigs, the same style of arithmetic as the worked example on the Business model page. Money is committed to a gig only when a creator accepts, and released only as views are read on day 3, day 7 and day 14, so the fund is never spent faster than the product delivers.

3. Preset G, "Creator story"

Every gig has exactly 10 checklist items. The platform review marks each one pass or fail, and all 10 must pass before the post goes live. Brands pick one of the preset checklists described on the Product spec page. Preset G is used only by dogfood campaigns. Items 1, 2, 3 and 5 are specific to the creator story. Items 4, 6, 7, 8, 9 and 10 are the standard items that appear in every preset. Here is the full list.

1. Show the offer screen or the wallet screen in the first 3 seconds.
2. Say "Distribution Club" out loud once.
3. Say in your own words how a gig works (offer, make, upload, posted for you, paid as views come in).
4. Video is vertical (9:16), between 20 and 60 seconds (assumption: the brief sets this exact range for Preset A only and says every preset narrows the 15 to 90 second upload window; this page uses the same range for Preset G).
5. The call to action is "join with my link in the caption".
6. The caption contains the campaign hashtag and tags the brand's handle.
7. The creator appears on camera or narrates; all footage is original (no reused clips, no stock).
8. No competitor product or brand is visible or mentioned.
9. Audio is clear and music is from the platform's licensed library only.
10. Paid partnership disclosure is present ("#ad" plus the platform's paid-partnership or branded-content label). The platform adds this automatically at posting; the creator must not remove it.

For brand zero, "the brand's handle" in item 6 is Distribution Club's own account on that platform, and the caption also carries the creator's referral link (section 7). On a platform where a caption link cannot be tapped, the link is written out as text and the creator may also put it in their profile (assumption).

A good creator-story video is small and specific. The creator holds up the phone, shows the offer or the wallet, and says what happened to them: an offer came in one evening, I accepted, I filmed the next morning, I uploaded it, the platform reviewed it and posted it for me, and I was paid as the views came in. It can say what was annoying. A creator may say the review took a day, or that an offer expired before they saw it. Honest opinion is required under the disclosure regimes we follow worldwide (the US FTC Endorsement Guides among them), and the checklist never asks for an opinion, only for facts: say the name, show the screen, explain the steps.

Three things are not allowed. No earnings promises: a creator may show their own real wallet but must not say "you will make this much a month" or anything like it. No fake screenshots: every screen shown must be the creator's own account, captured on their own phone. No scripted praise: the video is the creator's own words, not a script from us. A submission with an earnings promise or a screen that is not the creator's own fails review and gets the normal single revision round with a 48-hour deadline (assumption).

4. The founder as creator, brand and ops

Distribution Club is one person through 2027 (brief section 9). The founder has a real creator account with at least one connected social account, sets availability hours, and completes at least 1 gig a week from day 13 of the MVP sprint through the whole of Phase 1 build-out and all of Phase 2, acting as creator, brand and ops all at once. The accountant and the lawyer are paid per job and are not asked to take gigs (assumption). Alongside the founder, 20 friendly creators recruited by hand (brief section 15) act as the creator side of dogfooding at real volume, each with a real creator account, at least one connected social account and availability hours set, completing dogfood gigs on the same terms as any other creator. The founder's account and the 20 friendly-creator accounts must pass the same eligibility rules as everyone else: age 18 or over, a median of at least 5,000 views per post over the last 90 days on a connected account, at least 10 public posts in the last 90 days, account age at least 90 days. A friendly creator who cannot pass the median views rule is marked by ops as a test account, which receives dogfood offers only (assumption). Like every other creator, the founder and each friendly creator post the approved package themselves once a submission is approved and paste the link for the platform to check, so these gigs also exercise the link-check step, not just the offer, review and payout steps.

Founder and friendly-creator earnings are real. Any of them who completes a gig earns pay based on real views: a floor of 20% of expected earnings released at publish, then the rest released as views are read on day 3, day 7 and day 14, up to the maximum payout of expected earnings, unless a bonus is switched on, held as Pending for 48 hours, then Available, then paid to their own bank account on the 1st if the balance is at least $20. They go through the same payout setup as any creator: identity verification, the tax details the payout rail asks for (PAN for creators in India) and a bank account, UPI ID or Payoneer account through the payments partner. The accountant decides how the founder's own payouts are reported at year end, since the founder is the business owner rather than an independent contractor like the other creators (assumption). The reason for insisting on real money is simple: a payout that fails for the founder or a friendly creator on 1 October 2026, the first payout run, is a payout that would have failed for a stranger, and the founder is the one who sees the error.

The founder's account and the 20 friendly-creator accounts are excluded from the public creator counts, from any completed-gig counts we publish, and from the metrics in section 11. These accounts receive dogfood offers only, never offers from pilot brands, so no brand ever pays for the founder's or a friendly creator's own content (assumption). The founder built the brand dashboard and reviews these videos with the same 10 items and the same one-line reasons as anyone else's.

5. First-gig guarantee

5.1 The rule

Every new creator who passes eligibility gets their first offer within 7 days. If no brand campaign matches them inside 7 days, a dogfood campaign gig is dispatched to them. An offer counts whether the creator accepts, declines or lets it expire; the promise is that the phone buzzes within a week. The guarantee applies from the first day dispatch is real, not from some later date once the beta opens: the first creator eligible under it is the founder, a friendly creator, or the first pilot brand's creator on day 13 or 14 of the MVP sprint (19 or 20 September 2026), and every creator who signs up after that gets the same 7-day promise, backed by the dogfood campaign that has run every week since.

5.2 Why it matters

Sign-up is not the goal. A completed gig, a wallet balance and a first payout are the goal. A creator who signs up, connects an account, sets availability hours and then hears nothing for two weeks concludes that the product is empty and deletes the app. The beta target is that 60% of new creators complete a first gig within 14 days. An offer inside 7 days, a 72-hour default deadline, a review decision inside 48 hours and posting straight after approval add up to a first post inside 14 days for most creators who accept.

5.3 How it is implemented

A daily job finds every creator who passed eligibility more than 5 days ago and has not yet received any offer, from any campaign. It adds them to the candidate list for the next wave of the current week's dogfood campaign, using the dogfood campaign controls in the ops console (assumption). Day 5 is chosen because dispatch waves run for up to 48 hours: starting on day 5 means the offer lands by day 7 even if the first wave does not reach the creator. The offer still obeys the normal rules: it is sent only inside the creator's availability hours, it expires after 30 minutes, and it counts towards the limit of 3 active gigs. The creator sees "Distribution Club campaign" as the brand, the CPM and expected earnings for their league, whether a bonus is on and the deadline, exactly like any other offer.

So that any new creator can match, the weekly dogfood campaign is set up with all eight leagues (Bronze 1, Bronze 2, Silver 1, Silver 2, Gold 1, Gold 2, Ruby, Diamond), the ten V1 categories (AI tools, SaaS and web apps, developer tools, productivity apps, fitness apps, health apps, lifestyle apps, education apps, kids apps, games), matched by audience rather than by creator location, and no city filter (assumption). If more creators are waiting than the week's slots can take, the oldest sign-ups go first and the founder adds funds to the campaign from the same weekly budget (assumption). Creators who did receive a brand offer in their first 7 days are not touched by this job.

5.4 What it costs

A first gig's cost depends on the creator's own median views, since pay is a CPM per 1,000 views rather than a fixed fee. Take Marcus, a Bronze 1 creator with a 6,000-view median (the floor of the eligibility range) and a Tier 1 audience posting on Reels for a SaaS and web apps campaign, so his CPM is $5.00 ($2.50 base, Reels 1.25, category factor 1.6, Bronze 1 1.00): his expected earnings are $30.00 (6,000 views, in thousands, times $5.00), so his gig takes $40.00 from brand zero's fund (expected earnings plus the $2 running cost, divided by 0.8). Of that $40.00, $30.00 is expected creator pay, $2.00 is the running cost paid from the pool before creators, and $8.00 corresponds to the platform's 20% cut, already recognised as revenue when brand zero's deposit landed, money moving from one of our accounts to another, so it is not a new cost. The real cost is the creator payout: a $6.00 floor (20% of expected) released the moment the post is live, then whatever real views add up to at the day 3, day 7 and day 14 reads, up to the $30.00 maximum payout (his expected earnings). Marcus reaches 3,000 views by day 14 and earns $15.00 in total, so $15.00 of the $30.00 reserved returns unused to brand zero's creator pool. At a typical Bronze 2 median (12,500 views, same audience) a first gig expects $62.50 and takes about $80.63 from the fund; at a typical Gold 1 median (75,000 views, league factor 1.05) it expects $393.75 and takes $494.69; at a typical Diamond median (750,000 views, league factor 1.10) it expects $4,125.00 and takes $5,158.75, from the same rate card. The running cost, review, payout transfer, storage and the card fee's share, is inside every one of these figures and is never itemised, even to brand zero.

6. Payout day as a monthly event

On the 1st of every month the founder runs the payout batch. Every creator with completed payout setup and an Available balance of at least $20 is paid to their verified bank account; below $20 rolls over. The same day, brand zero funds a dogfood campaign called "Payout day". Its candidate list is limited, through the ops console dogfood controls, to creators who were paid in that day's run (assumption). The offer asks for a short video showing the payout notification. Amounts may be hidden; the creator chooses. Preset G applies unchanged: the wallet screen in the first 3 seconds satisfies item 1, and the video explains how a gig works, says "Distribution Club" once and ends with the referral link call to action.

The first payout run is 1 October 2026 (brief section 28), paying dogfood earnings from day 13 of the sprint and the first pilot brand's earnings from day 14 onward; it runs monthly after that, on 1 November and 1 December 2026, 1 January 2027, and on through 1 February, 1 March, 1 April, 1 May and 1 June 2027 and every month beyond. The "Payout day" campaign is the dogfood campaign for that week and is paid from that week's budget, not from extra money (assumption).

A monthly rhythm helps in three ways. It gives the content a natural beat: every month there is a fresh group of creators who were just paid and have a reason to say so. It gives new creators proof, in another creator's own words and on that creator's own account, that the product pays. And it gives the founder a fixed date to watch the whole money path end to end, from a floor or view-read release through the 48-hour hold to a bank transfer, and to see the failures in one place. If the 1st falls outside a creator's availability hours, the offer waits for the next window; this is normal dispatch, not a special path.

7. Referral loop

Each creator has a referral link. When a creator who signed up through that link completes a first gig, the referrer earns $5 into their wallet. "Completes" here means the referred creator's first post is live and the platform has captured the post ID, the moment the floor releases on their first gig. The reward is paid at that moment, not at sign-up, so a referral that never leads to a gig earns nothing.

The link appears in the caption of every dogfood video, because Preset G item 5 makes the call to action "join with my link in the caption". It also appears in the creator app, so a creator can send it to a friend directly. During Phase 2 the beta is private; a referred creator who passes eligibility is admitted to the beta pool (assumption). From Phase 3 the referral programme is open to all creators, as listed on the Rollout plan page. So that we can tell a sign-up from a dogfood post apart from a sign-up from a direct share, each dogfood post carries its own variant of the creator's link (assumption).

Abuse controls, so that the $5 is not worth gaming:

• One payout account per person. A person cannot refer a second account of their own and collect twice; identity verification at payout setup enforces this.
• Device and account fingerprinting. A referred account created on the same device as the referrer is flagged for ops review.
• The reward sits as Pending for the same 48-hour fraud hold as any other release, while the automated checks run on the referred creator's first post.
• No reward for the founder's account or the 20 friendly-creator accounts, whether that account is the referrer or the referred creator (assumption).
• The reward is paid only when the referred creator's first post is confirmed live and its floor releases, so bought sign-ups that never do a gig cost nothing.
• Referral rewards do not affect the creator score, which counts only gigs.

8. Friction capture

After every dogfood gig, the app asks the creator five questions. Each is answered 1 to 5, where 1 is bad and 5 is good.

1. Offer clarity. Was it clear what the gig asked for, what it paid, and by when?
2. Upload ease. How easy was it to upload the video and submit it?
3. Review feedback quality. If changes were requested, was it clear what to fix?
4. Posting trust. How comfortable were you with the platform posting on your account?
5. Payout clarity. Did you understand what was Pending, Available and Paid, and when the money would arrive?

Then one free-text box: "Anything confusing, broken or slow?" The survey is asked after dogfood gigs only, not after brand gigs, so creators working for pilot brands are not interrupted.

A $10 bug bonus is paid for each confirmed bug reported by a creator; it is not called a "credit": the site does not use that word for money at all. A bug is something the product does that it should not do: a crash, a wrong number in the wallet, an offer that cannot be opened, a post that shows as live when it is not. A request for a feature, or a disagreement with a rule such as the 30-minute offer expiry, is welcome but is not a bug. Ops confirms a bug by reproducing it on a test device, or by the founder finding it in the error tracking logs. The first creator to report a distinct bug gets the bonus; later reports of the same bug get a thank-you (assumption). The bonus lands in the wallet as Available and is paid on the normal cycle (assumption).

Every week the founder reads all survey answers and bug reports and picks the top 3 fixes for the next release. Phase 2 has weekly releases, so a bug reported on Monday can be fixed for creators the following week. The five scores are tracked over time as a dogfood metric (section 11); the free-text answers are the raw material for the next week's priorities.

9. Budget and cost per creator

9.1 Budget by phase

The two-week MVP moves dogfooding forward from the old plan's Phase 1 weeks 13 to 14 (11 to 22 January 2027, 2 weeks) to day 13 of the sprint (19 September 2026), which adds 1 week in the sprint itself plus 18 weeks of Phase 1 build-out, 19 weeks in total at $2,000 instead of the old 2, before the beta even opens. That is 17 extra weeks: 17 times $2,000 is $34,000 more than before. The $2,000-a-week period now runs 28 weeks in total (1 plus 18 plus the unchanged 9 weeks of Phase 2), which is $56,000. Phase 3 is unchanged at 13 weeks and $65,000. The new grand total is $56,000 plus $65,000, which is $121,000 across 41 weeks, up from the old plan's $87,000 across 24 weeks.

These totals are the fund, the money that can reach creators, charged to brand zero's card as one deposit, the same as any brand pays: $2,000 a week from the MVP sprint through Phase 2 and $5,000 a week in Phase 3. 20% of each deposit becomes the platform's revenue at once, and the other 80% is the creator pool the week's gigs draw from. The cash that really leaves the company is: creator earnings released, collection cost of about 1.1% on the deposit (about $22 a week on $2,000, about $55 a week on $5,000), referral rewards of $5 each, and bug bonuses of $10 each. Creator pool money that is not released in the 14-day window returns as fund balance and is used the following week. None of this spend is fixed or promised: it is a choice the founder makes from profit, and it is not budgeted into the Scenarios page's model, which uses a cost base of no salaries, $2,500 a month of other fixed costs, and no dogfood spend (brief section 35). If the founder chooses not to run dogfooding at the budget shown here in a given week, or at all, the Scenarios numbers are unaffected either way.

9.2 Cost per eligible creator sign-up

The target is a cost per eligible creator sign-up from dogfood campaigns of under $15 by the end of Phase 2 (26 March 2027). "Eligible" means the creator passed all the eligibility checks, not merely installed the app. Here is the arithmetic for one Phase 2 week, with the assumptions stated.

• $2,000 a week buys about 31 Bronze gigs on the Bronze auto split (section 2).
• Assume each dogfood post averages 1,500 views in its 14-day window (assumption). That is 31 x 1,500 = 46,500 views a week.
• Assume 1 eligible sign-up per 400 views (assumption). That is 46,500 divided by 400, about 116 eligible sign-ups a week.
• $2,000 divided by 116 is about $17.24 per eligible sign-up, above the $15 target: software gigs pay creators 60% more than the old food example for the same views, so the same weekly pool buys fewer gigs and fewer views than the old arithmetic assumed (assumption).

At the 1,500-view assumption, the dogfood budget lands slightly above the $15 target in both phases: about $17.24 per eligible sign-up in Phase 2 and about $17.30 in Phase 3. This is because a software gig at category factor 1.6 pays a creator 60% more than the old food example for the same views, so the same weekly pool buys fewer gigs and fewer views. There are three ways to close the gap: accept a cost per eligible sign-up of about $17, raise the per-post view assumption above 1,500, or shift some of the dogfood budget to Bronze 1 only, where the lower league factor buys more gigs per dollar. The founder's default is the first option, accepting about $17 per eligible sign-up rather than changing the view assumption or the league mix, and the whole comparison is marked as an assumption.

The break-even is easy to remember: $2,000 divided by $15 is about 134 sign-ups a week, which at 1 per 400 views is 53,600 views a week, or about 1,729 views per post across 31 posts, above the 1,500-view assumption, which is why the Phase 2 number above lands above target at that assumption. In Phase 3, $5,000 a week buys about 77 Bronze gigs (as above) and needs 334 eligible sign-ups a week to stay under $15, which at 1 per 400 views is 133,600 views a week, or about 1,735 views per post across 77 posts, also above the 1,500-view assumption. At the 1,500-view assumption itself, 77 posts give about 289 eligible sign-ups (77 x 1,500 = 115,500 views, divided by 400), and $5,000 divided by 289 is about $17.30 per eligible sign-up, so the Phase 3 cost per sign-up also lands above $15 given these assumptions (assumption). The table shows how the number moves.

Counting rules. The metric counts the whole weekly campaign budget, even though earnings short of the expected amount come back, because that is the conservative way to count. Referral rewards and bug bonuses are tracked as separate lines, not inside this number (assumption). The MVP sprint and the first weeks of Phase 1 build-out are excluded, since only the founder, the friendly creators and the earliest pilot brand's creators are posting then and the number would mean little. The metric is measured over the whole of Phase 2 on 26 March 2027 and monthly after that.

10. Guardrails

Dogfooding can quietly make the marketplace look better than it is. These rules stop that.

• The 30% cap. Dogfood gigs never exceed 30% of any creator's completed gigs in a 30-day period, so the creator score (0 to 100, built from reliability, quality, performance and responsiveness) reflects real brand work. At each dispatch, a creator who would go over 30% is skipped for the dogfood offer. The first-gig guarantee gig, a creator's very first gig, is exempt, because without the exemption the guarantee could not work; everything after it, including "Payout day" gigs, is inside the cap (assumption). The founder's account and the 20 friendly-creator accounts are exempt from the cap because they receive dogfood offers only and are excluded from every public number (assumption).
• Labelling. Dogfood campaigns are labelled "Distribution Club campaign" in every dashboard: on the creator's offer, in the ops review queue, in reports and in the CSV export. Nobody should ever have to guess whether a gig was brand-funded.
• Excluded from brand-funded volume. Dogfood campaigns are left out of brand-funded volume in any reporting. Brand zero is not counted as a paying brand, its deposits do not count towards the Phase 3 target of $200,000 in monthly fund deposits, and dogfood gigs are reported on their own line next to the target of 2,000 completed gigs a month, never mixed in.
• Same content rules. Dogfood campaigns follow the prohibited-content rules and the disclosure rules like any other campaign. Every dogfood post carries "#ad" and the platform's paid-partnership or branded-content label. Brand zero is not in any prohibited category and must not look like one; see section 13 on the get-rich-scheme risk.
• Founder and friendly-creator gigs excluded from public counts. The founder's account and the 20 friendly-creator accounts (21 in total) are excluded from the public creator count, the completed-gig count and the metrics in section 11.
• Dogfood distortion is a named risk. Too many platform-funded gigs can make the marketplace look healthier than it is and inflate creator scores. It is risk 13 on the Challenges page, and section 13 below says what we do about it.

11. Metrics and targets

These are the numbers the founder looks at in the weekly review. Two of them (30% of sign-ups from dogfood and referral, 60% first gig within 14 days) are also beta success criteria, so the beta is judged on them at the end of Phase 2.

The last two rows are not in the brief's metric list. They are added here because they are the two rules on this page that can break without anyone noticing.

12. Week by week plan

Weeks run Monday to Sunday, except where the MVP sprint's own days fall outside that pattern. Day 13 of the sprint (the first dogfood campaign) is Saturday 19 September 2026 and day 14 (the first pilot brand) is Sunday 20 September 2026; Phase 1 build-out runs from Monday 21 September 2026 to 22 January 2027; Phase 2 starts on 25 January 2027; Phase 3 starts on 29 March 2027 and ends on 25 June 2027. The rows before the beta are grouped by month to keep the table readable; from the beta onward they follow the same week blocks as before. Full phase dates are on the Timeline page.

13. Risks of dogfooding and how we handle them

Inflated marketplace health

In Phase 2 the dogfood budget could fund about 31 gigs a week, while the brand target is 150 completed brand gigs over 9 weeks. Without care, dogfood gigs would outnumber brand gigs and make fill rate, completed gigs and deposits look like something brands are paying for. We handle this with the labelling and the exclusion from brand-funded volume in section 10, and by measuring the beta success criteria (fill rate, on-time submission, first-pass approval and the rest) on brand campaigns only (assumption). The 30% cap means much of the Phase 2 dogfood budget may go unspent. That is fine: the budget is a ceiling, and unspent fund rolls over.

Creators seeing the platform as the only brand

If most of a creator's gigs come from brand zero, they learn the wrong lesson: that the platform pays, rather than brands. The 30% cap keeps dogfood a minority of every creator's work after their first gig. Preset G item 3 makes the creator explain that a gig comes from a brand's campaign. And the dogfood offer looks like every other offer, with a brand name, a CPM and expected earnings, and a deadline, so the habit it builds is the habit of taking brand gigs.

Spending the budget on views that do not convert

Views are not sign-ups. The cost per eligible sign-up is reviewed weekly. If it is over $15 for 3 weeks in a row, the founder lowers the weekly fund rather than raising it, and the founder changes the brief and the call to action before spending more (assumption). The downside is bounded: fund is committed only when a creator accepts and released only as views are read, so a week of bad videos costs at most that week's budget.

Guarantee outrunning the budget

In Phase 3 the weekly dogfood campaign has about 77 slots. If more creators than that pass day 5 without a brand offer in a single week, the guarantee cannot be met from the dogfood budget alone. Ops watches the waiting list daily, brand campaigns must carry most first gigs, and the founder can add funds within the weekly budget (assumption). If the list keeps growing, the fix is more brand campaigns in the categories where creators are waiting, not more dogfood.

Founder and friendly-creator bias

The founder and the 20 friendly creators know where every button is and forgive what a stranger would not. That is why dogfooding runs on these 21 accounts from day 13 of the sprint before any outside creator depends on the product, why survey answers from these accounts are kept apart from other creators' answers (assumption), and why the founder runs the payout batch by hand and reads every free-text answer.

Creator-story videos looking like ads for a get-rich scheme

A video of a person waving a phone and talking about money can look like a scam. Preset G asks how a gig works, not how much you can make. Earnings promises fail review. Amounts may be hidden on "Payout day". The referral reward is $5, small by design. Every video carries "#ad" and the platform label. Honest opinion is required, so creators can and will say what was slow. And there is nothing to upsell: creators pay no fee at MVP, there is no paid priority tier, and ranking is never sold, so the video has no product to push except a real gig.