Tier	Piece	Status	Round	Verdict	Remaining gap
T1	1. Landing page, land plus build Landing page	superseded	1	Ranking c > b > a Winner c: Variant c is the only entry where the hero states the offer and the next step in the same glance and the first click (pick an area) immediately shows county-record detail for that area, with no real lot address on the page and no forced SMS consent. c: FAIL T4; b: FAIL T4; a: FAIL T4	Variant c needs its colours cut back to exactly the four brand colours plus #F6F6F7 and #54585A, the logo raised to at least 96px of visible mark, and its 'the builder pays our fee' line corrected so it does not imply the builder also pays for lot-side representation.
T1	1. Landing page, land plus build Landing page	judged	2	Ranking b > c > a Winner b: Variant b is the only entry that answers this buyer's real questions in weeks, dollars and decisions (a week-by-week calendar, a who-pays-whom table, one-time close versus draws) while still putting the offer and a complete seven-field form with one-time close options in the hero. b: pass; c: pass; a: pass	Variant b still needs the speed-to-lead line changed to the brief's 5 minute / 9am promise and the [LOT-SIDE COMPENSATION TERMS TO BE SET BY JACOB] placeholder filled so 'What does this cost me?' is complete on both commission events.
T1	2. Landing page, BOYL Landing page	superseded	1	Ranking a > b > c Winner a: Variant a is the only entry that shows the actual one-page report on a real county parcel right under the form, so the lot owner sees exactly what their lot will get before handing over four fields, with the builder-pays line beside the button. a: FAIL T4; b: FAIL T4; c: FAIL T4	Variant a needs the logo set to at least 1in of visible mark and every non-palette neutral or tint swapped for #F6F6F7, #54585A, white or the four brand colours; after that it is ready for Jacob's review.
T1	2. Landing page, BOYL Landing page	superseded	2	Ranking b > a > c Winner b: Variant b is the only entry that tells the lot owner what each of the four answers does to a builder's bid and then puts the loan-type decision (one-time close at 45 to 60 days vs cash or draws at 12 to 18 months) in front of them, so it sells the report and qualifies the lead in the same scroll, with a separate SMS consent box and the builder-pays line under the button. b: FAIL T4; a: FAIL T4; c: FAIL T4	Variant b needs its merge-field chips set in Poppins, #E3E4E5 swapped for #F6F6F7 or #54585A, and the mobile logo kept at 1in or larger, and on phones the form should move up so the lot address field shows on the first screen instead of about 940px down.
T1	2. Landing page, BOYL Landing page	judged	3	Ranking b > c > a Winner b: Variant b is the only variant that tells the owner what each of the four answers does to a builder's price and surfaces the one-time-close versus draws decision, so it talks in decisions and dollars while still passing the three-second hero test and every compliance line. b: pass; c: pass; a: pass	All three show the report as a table of raw {{merge}} fields, so no visitor ever sees what a filled-in report on a real-looking lot reads like, which is the thing that would make the offer tangible.
T1	3. Email nurture, land plus build Emails 1 to 3 (vs existing 01 to 03)	superseded	1	Ranking a > b > existing Winner a: Variant a is the only entry whose subject and opening line put county parcel data for the reader's own target area up front (vacant lot count, flood zone, permits within a mile), with feasibility as the whole content, in correct brand colours, compensation language and contact details. a: FAIL T4,T2; b: FAIL T2; existing: FAIL T5,T6,T2	Variant a must replace the Courier New block in email 3 with Poppins and rewrite email 3's subject and first line around a parcel field (for example deed restrictions or permits within a mile of {{target_area}}) before Jacob can approve it.
T1	3. Email nurture, land plus build Emails 1 to 3 (vs existing 01 to 03)	judged	2	Ranking a > b > existing Winner a: Variant a is the only entry whose emails each open with a county-record fact about the reader's own target area (vacant lot count, absentee owners, flood zone, a real platted legal description) and teach one feasibility decision per email, which is exactly what this persona needs, inside a branded layout. a: pass; b: pass; existing: FAIL T5,T6	Variant a's emails 2 and 3 need a fallback for leads with no chosen lot (target-area figures in place of lot-level {{flood_zone}} and permit fields), and the header mark at 64px is under the 1in logo minimum.
T1	3. Email nurture, land plus build Emails 4 onward, one sequence (vs existing 04 to 12)	superseded	1	Ranking a > existing Winner a: Variant a covers every required topic in nine on-brand, answer-first emails with parcel merge-field boxes, compliant compensation language, Jacob-set placeholders and a full schedule through the month-24 evergreen loop, while the existing asset fails three tests outright. existing: FAIL T2,T5,T6; a: FAIL T2	Variant a needs a parcel merge field in each subject and first line (for example the shortlist lot's address with its flood zone or MUD rate) so that it passes T2, and emails 13 to 17 in its schedule are still unwritten.
T1	3. Email nurture, land plus build Emails 4 onward, one sequence (vs existing 04 to 12)	judged	2	Ranking a > existing Winner a: Variant a is the only entry that clears every test: every subject and first line is built from the lead's own parcel fields (flood zone, sewer on record, MUD, permits within a mile), it has a full send-week schedule that runs into a quarterly evergreen loop to month 24, it uses only the exact brand hexes and Poppins, it uses [RANGE TO BE SET BY JACOB] where a cost belongs, and its footer states the builder-pays framing correctly. a: pass; existing: FAIL T2,T5,T6,render	Cut each bold lead in variant a to one short answer sentence followed by the detail, and have Jacob confirm the live parcel fields (flood_zone, sewer_on_record, mud_tax_rate, nearest_permits_1mi) exist and that the send-hold rule for empty fields works in the ESP.
T1	4. Email nurture, BOYL Emails 1 to 3	superseded	1	Ranking a > b Winner a: Variant a turns the parcel record into a lot-by-lot report and a site-work checklist keyed to the recipient's own flood zone, water and septic fields with no invented numbers, while variant b slips an unsourced dollar claim into Email 2. a: pass; b: FAIL T6	Variant a still needs the invented 12-to-24-month claim cut, and Email 3's subject and first line should use a fact from the lot record (flood zone or acres) rather than generic financing copy.
T1	4. Email nurture, BOYL Emails 1 to 3	judged	2	Ranking a > b Winner a: Variant a turns the parcel record into the content of every email (seven lot fields in email 1 with what each one decides, a site work checklist keyed to the lot's acres and flood zone in email 2, and the one-time close vs draws choice in email 3), while variant b's parcel data stops after email 1. a: pass; b: pass	Cut or source the '12 to 24 months from building' claim in email 1 and rewrite email 3's first line around the reader's own lot (its flood zone or lot equity toward the down payment) so all three subject plus first lines pass T2 on their own.
T1	4. Email nurture, BOYL Emails 4 onward, one sequence	superseded	1	Ranking a Winner a: Variant a is the only entry and it covers every required topic (cost range placeholder, site work, one-time close vs draws, MUD, septic with Montgomery County OSSF, builder selection, weeks) with answer-first openers, correct fee language and no em dashes, but it still fails T2. a: FAIL T2	Rewrite all seven subjects and opening sentences to lead with the recipient's own parcel facts (flood zone, MUD district, utilities on record, permits within a mile), and while there drop the off-palette #E3E4E5 and #FFF6D6 tints and the duplicated compensation paragraph in email 9.
T1	4. Email nurture, BOYL Emails 4 onward, one sequence	judged	2	Ranking a Winner a: Variant a is the only entry and it passes T2, T5 and T6 across all seven emails: every email opens with the answer, covers its required topic (cost range placeholder, four site work lines, one-time close vs draws, MUD, conventional vs aerobic septic with OSSF permit, builder selection, timeline in weeks), carries the correct builder-pays fee line, and puts a lot-specific county field in the subject. a: pass	Before Jacob can approve, the county layer has to be joined so flood zone, utilities, MUD and permit counts actually populate, and email 10's 10 to 12 month headline has to be reconciled with its 41 to 75 week table (and loan_type moved out of the county record box), with the small raster logo in a white tile replaced by the primary mark on navy at legible size.
T1	5. SMS sequences SMS, land plus build (vs existing sms.md)	superseded	1	Ranking a > b > existing Winner a: Variant a is the only entry where the texts themselves carry parcel-level facts about the lead's own target area or lot (live lot count, FEMA zone, permits within a mile), with a compliant opt-out first message, the builder-pays line, verified character counts, and a clean branded table plus phone previews. a: pass; b: FAIL T2; existing: FAIL T5,T6,T2	Variant a's opening text needs a real parcel fact for the lead's target area instead of a promise to pull one, and the permit and flood-zone merge fields need a confirmed data pull before any of those texts can send.
T1	5. SMS sequences SMS, land plus build (vs existing sms.md)	judged	2	Ranking a > b > existing Winner a: Variant a makes every one of its 12 texts point at something from the county parcel record (flood zone, utilities, permits within a mile, lot size), keeps all of them at or under 160 characters with accurate counts, and opens with sender, brokerage and opt-out in one message, while variant b only uses the parcel data in 4 of its 13 texts. a: pass; b: pass; existing: FAIL T5,T6,T2	Jacob still has to approve the live merge sources variant a depends on ({{lots_listed}}, {{nearest_permits_1mi}}, {{alt_area}}) and give text 0 a reply hook, since without those feeds the parcel-led texts cannot go out as written.
T1	5. SMS sequences SMS, BOYL	superseded	1	Ranking b > a Winner b: Variant b is the only entry with no automatic fail: every text is one segment, the first touch names Builder Connect, CB&A and STOP, loan type lands in text 2, parcel facts drive texts 3, 5 and 8, and the hot-feed opener is held for counsel. b: pass; a: FAIL T4	Text 1A should put one real parcel fact (acreage or flood zone) in the opener and text 3 needs its merge output fixed to read "FEMA Zone X" before Jacob approves it, and counsel still has to clear the hot-feed consent route.
T1	5. SMS sequences SMS, BOYL	judged	2	Ranking a > b Winner a: Variant a puts a parcel fact (flood zone, acreage, nearby permits, the cost placeholder) into nearly every text and its first message carries the full compliance set (sender, brokerage, STOP, HELP, msg and data rates), so it does the one thing competitors cannot copy more often than variant b. a: pass; b: pass	Variant a needs its 2 segment texts cut under 160 characters and text 4 rewritten to ask about water and sewer without inferring septic from acreage before Jacob can approve it.
T2	6. Meta ads Ad concepts, land plus build	superseded	1	Ranking a > c > b Winner a: Variant a is the only entry that pairs a true, sourced county fact (60,328 vacant residential lots on the Montgomery County roll) with a lot-report creative and an on-image call to action, so it is the clearest use of the parcel-data angle no competitor can copy. a: FAIL T4; c: FAIL T2,T4; b: FAIL T2,T4	Variant a needs its #E3E4E5 rules swapped to an approved neutral, and its literal {{flood_zone}} style placeholders replaced with real per-area county values (or a SAMPLE-marked filled lot), because a static Meta image cannot merge and would show raw curly braces to every viewer.
T2	6. Meta ads Ad concepts, land plus build	superseded	2	Ranking b > a > c Winner b: Variant b is the only creative that leads with a hard, true county fact (60,328 vacant Montgomery County lots, parcel data on each) and backs it with the buyer's actual timeline in weeks and days plus a phone number, so it reads as data no competitor has rather than a promise. b: pass; a: pass; c: FAIL T2,T1	Replace variant b's three merge-field placeholders with real values for each ad set's target area (or drop the box), and tighten the empty band in the 1080x1350 version, before Jacob can approve it for spend.
T2	6. Meta ads Ad concepts, land plus build	superseded	3	Ranking c > b > a Winner c: Variant c is the only one that ships as a finished ad: a named Willis buyer, a plain instruction to check flood zone and utilities before offering, real county counts with no template braces, and a visible call or email next step. c: pass; b: FAIL T1; a: FAIL T2	No variant yet tells a land-plus-build buyer a real parcel-level fact about the area they are shopping (for example how many on-market lots there sit outside the floodplain with utilities at the road), which is the one angle competitors cannot copy.
T2	6. Meta ads Ad concepts, land plus build	superseded	4	Ranking b > c > a Winner b: Variant b is the only creative a land buyer reads in three seconds as a complete plan: a hard county fact (60,328 vacant Montgomery County lots, parcel data on each), the four things checked before purchase, the lot closing in 30 to 45 days, and one phone call as the next step. b: pass; c: pass; a: FAIL T2	No variant yet shows a land-plus-build buyer a parcel-level fact about the area they are shopping, such as how many lots in that submarket sit outside the floodplain with utilities at the road, which is the one angle a competitor without the parcel data cannot copy.
T2	6. Meta ads Ad concepts, land plus build	judged	5	Ranking b > a > c Winner b: Variant b is the only one that answers both three-second questions at a glance (parcel-checked land search into a builder match, then a full-width call button with the number) while laying the process out in weeks and decisions exactly as the voice guide asks. b: pass; a: pass; c: pass	No concept yet shows a real, area-level parcel fact (for example share of acre-plus lots in a named area outside the floodplain) on the creative, which is the one thing that would make the parcel-data angle impossible for a competitor to copy.
T2	6. Meta ads Ad concepts, BOYL	superseded	1	Ranking a > b > c Winner a: Variant a is the only concept that makes the ad about the viewer's own lot, with a county-record report card showing flood zone, utilities and permits within a mile, so it is the only one a competitor without parcel data could not run. a: FAIL T4; b: FAIL T2,T4; c: FAIL T2,T4	Variant a needs its off-palette greys swapped for white or approved #F6F6F7/#54585A, and the literal {{merge_field}} braces replaced with a clearly labelled SAMPLE lot (or a confirmed data source) so the static image does not read as broken in the feed.
T2	6. Meta ads Ad concepts, BOYL	superseded	2	Ranking c > b > a Winner c: Variant c is the only entry where the offer (the report on your lot) and the owner's own county record (acres, street, flood zone, utilities, permits within a mile) both carry the creative, framed by a concrete site-work checklist that turns the report into a decision tool. c: pass; b: pass; a: FAIL T2	Jacob has not yet seen variant c filled with one real, confirmed county record (labelled SAMPLE), and variant c has no stated delivery plan for the per-lot merge fields (custom audience segments or a dynamic catalog), because Meta cannot personalise a static image for each owner.
T2	6. Meta ads Ad concepts, BOYL	superseded	3	Ranking a > c > b Winner a: Variant a is the only concept whose rendered ad is true as it will actually run on Meta: a filled, clearly labelled SAMPLE lot report (flood zone, water and sewer, permits within a mile) shows the offer in one glance, with one headline, one button and the builder-pays line, all in exact brand colours and Poppins. a: pass; c: FAIL T2; b: FAIL T2	No BOYL concept puts a true, verifiable county-level parcel fact on the creative itself; the winner shows a fictional sample, and the two that try for per-lot data rely on merge fields Meta ads cannot fill.
T2	6. Meta ads Ad concepts, BOYL	superseded	4	Ranking a > c > b Winner a: Variant a is the only concept where every claim on the creative is true as it will run: one headline about your lot, a clearly labelled SAMPLE report showing exactly what you get (flood zone, water and sewer, permits within a mile), one button, the builder-pays line, exact brand colours and Poppins. a: pass; c: FAIL T2; b: FAIL T2	No BOYL concept yet carries a county-level parcel fact that is both true and verifiable from the actual data pull (for example 25,051 individually owned vacant lots on the Montgomery County roll) alongside variant a's sample report, so the winner still proves the format, not the data.
T2	6. Meta ads Ad concepts, BOYL	judged	5	Ranking a > b > c Winner a: Variant a is the only concept where three seconds tells you exactly what you get and what to do: a headline about your lot, a clearly labelled SAMPLE one-page report showing flood zone, water and sewer and permits within a mile, and one button, all in exact brand colours, Poppins and the builder-pays line. a: pass; b: pass; c: FAIL T2	No BOYL concept yet pairs variant a's sample report with a true, verifiable county fact about owners like the viewer (for example variant b's 803 individual-owned lot starts a year against 25,051 owners), so the winner still shows the format, not the data.
T2	7. Postcards Postcard concepts (6x9 + 4x6, front + back)	superseded	1	Ranking c > a > b Winner c: Variant c is the only entry that places the full, uncropped logo at more than 1in wide with correct clear space, and it gives one offer (get the report on your lot) and one action (scan or visit builderconnect.net/lot-report) that read in three seconds. c: FAIL T4,T7; a: FAIL T4,T7; b: FAIL T4,T7	For variant c, move the return address to the upper left of the back, out of the address zone, and replace the CMYK builds in the print spec note with the guide's official ones (navy 88/70/60/65, yellow 0/25/100/0, blue 100/28/0/17, red 20/100/90/15).
T2	7. Postcards Postcard concepts (6x9 + 4x6, front + back)	superseded	2	Ranking a > c > b Winner a: Variant a puts the recipient's own flood zone, utilities and permit count on the front where it is seen first, which is the parcel-data angle no competitor can copy, and does it with the cleanest logo treatment and a complete print spec. a: pass; c: pass; b: FAIL T4	Variant a needs a verb-led call to action with a scannable QR or short URL on both sides, and the flood, utility and permit fields must be joined from real parcel data and the permit indicia set before Jacob can approve a print run.
T2	7. Postcards Postcard concepts (6x9 + 4x6, front + back)	judged	3	Ranking b > a > c Winner b: Variant b is the only card whose front headline is itself the parcel fact (homes permitted within a mile of your acres in your city), so with the logo covered it still tells the absentee owner something specific they did not know, and it closes with one clear phone action. b: pass; a: pass; c: pass	No variant has been proofed with a real sample parcel's data merged in, so fit, wrapping and truthfulness of the variable fields that carry T2 are still unverified.
T2	8. ISA call scripts Full ISA script with branches	superseded	1	Ranking a > b Winner a: Variant a is the only one an untrained ISA could run without asking a question: every box has an 'Anything else' row, the grading table is first-match and covers every lender/timeline combination, and it adds a pre-dial checklist, calling hours, inbound opening and exact GHL field names, while variant b leaves grading holes and loses glossary terms when printed. a: FAIL AUTO-typeface (T4: non-Poppins font); b: FAIL T3	Swap variant a's monospace field column to Poppins and fix its print layout, which currently spreads to 36 Letter pages with the text templates crushed into a narrow column, one per page.
T2	8. ISA call scripts Full ISA script with branches	superseded	2	Ranking a > b Winner a: Variant a is a true decision tree where every answer, including unlisted ones, wrong person, not a good time and do-not-contact, routes to a named next box with exact log fields, while variant b's cards leave unlisted answers unrouted and its print output pushes every card's footer onto a separate near-blank page. a: FAIL T3; b: FAIL T3	Variant a needs a fixed attempt count, plus definitions or real names for the overflow calendar, the Dow commercial line and the phone-number source, so a new hire can run every step without asking, and its SMS template table needs a wider text column so it is readable in print.
T2	8. ISA call scripts Full ISA script with branches	judged	3	Ranking a > b Winner a: Variant a is a true decision tree where every box has an exact line, an 'Anything else' row and one next step, plus a pre-dial checklist, a fixed 6-attempt cadence and an explicit 'Text held' rule, so a new ISA never has to choose between two conflicting instructions. a: pass; b: pass	Neither version settles where phone numbers come from (printed list only, or a manual search or skip trace), so the opening and the 'how did you get my number' answer stay unresolved until Jacob decides.
T3	9. Blog posts Blog post 1	judged	1	Ranking a Winner a: Only entry, and it earns the pass: answer in sentence one, five Montgomery County site-work line items, the correct builder-pays-the-fee language twice, every cost held as [RANGE TO BE SET BY JACOB], 847 words with a full SEO editor note, no em dashes and no banned words. a: pass	Jacob has to set six cost ranges before it can publish, and the merge-field box needs to become a static checklist or a filled SAMPLE lot (also soften 'costs you nothing to ask' and swap the off-palette #E3E4E5 divider for a brand gray).
T3	9. Blog posts Blog post 2	judged	1	Ranking a Winner a: Variant a is the only entry and it earns its place: the answer is in the first sentence, there are no em dashes or banned words, it states the builder-pays fee framing correctly, it quotes no rate or price, and the SEO kit (title, meta, target query, five question H2s, five-question FAQ) is complete at 1,115 words. a: pass	Rewrite the 'What does Builder Connect check' section as public copy (no merge fields, no internal list stats, no false 'most'), and add one real Montgomery County MUD example with a sourced rate and a weeks figure for the septic-permit saving so the post talks in dollars and weeks rather than 'many' and 'often'.
T3	9. Blog posts Blog post 3	judged	1	Ranking a Winner a: Variant a is the only entry, and it earns the spot: the answer is in sentence one, it gives a Montgomery County decision path in feet, permits and weeks, every cost is held as [RANGE TO BE SET BY JACOB], it states how the fee works, and the count is 1,157 words with no em dash or banned word. a: pass	Jacob has to set the seven [RANGE TO BE SET BY JACOB] cost ranges and swap the merge-field box for a real filled parcel example (or an address lookup) before the post has a single concrete dollar figure or a parcel proof a reader can see.
T3	9. Blog posts Blog post 4	judged	1	Ranking a Winner a: Variant a is the only entry and it passes T5 and T6: the answer is in sentence one, there are no em dashes or banned words, the builder-pays-the-fee disclosure is correct, and the only cost figure is the [RANGE TO BE SET BY JACOB] placeholder. a: pass	Jacob needs to set the site-work range and a lender needs to review the post, and it still needs at least one real Montgomery County process detail and a week-based timeline to rank for local search. The editor note says 950 words, but the body measures about 904.
T3	10. Social content 20 posts + 4 week calendar	superseded	1	Ranking a Winner a: Variant a is the only entry and it is a solid plan: all 20 posts are tagged by persona and platform, every post has its fee line, the calendar is complete with owners, and every county figure comes from the brief, but it does not pass. a: FAIL T4	Replace the monospace {{field}} chips on the 8 graphics with real county-level numbers set in Poppins, and rebuild the tiles to the guideline's social treatment (logo with proper clear space plus the chevron accent or four-colour strip) so the large empty navy area stops reading as an unfinished template.
T3	10. Social content 20 posts + 4 week calendar	judged	2	Ranking a Winner a: Variant a is the only entry and it clears every applicable test: 20 tagged posts with a complete Mon to Fri four week calendar naming Taylor, Philip, Rey and Jeremy, eight rendered 1080 PNGs in exact brand hex and Poppins, no em dashes or banned words, every post carrying the builder-pays fee line, and nine posts built on brief-sourced county parcel figures. a: pass	Jacob needs the graphic set redesigned so each tile fills its frame, uses the guide's social tile elements (four-colour strip, chevron accent) and actually matches its written visual direction, instead of eight near-identical navy cards with a dead lower half.
T3	11. Video scripts 3 lot walkthroughs + 1 build explainer	superseded	1	Ranking a Winner a: Variant a is the only entry, and its tables are complete: every row has timecodes, merge-field on-screen text, B-roll and lines Philip can read straight off, and the compensation line is right in every outro, but it still fails T3. a: FAIL T3	Name the actual lot for each walkthrough (or the rule for choosing one and who gets the owner's written consent), define the parcel pull step by step, and remove the address/parcel card wherever consent is missing; also fix walkthrough 1, which says '60 seconds' in a 75 second script.
T3	11. Video scripts 3 lot walkthroughs + 1 build explainer	superseded	2	Ranking a Winner a: Variant a is the only entry and it is close: four complete two-column shot/script tables within length, a clear release-and-parcel-pull rule set, correct buyer's-rep and builder-pays framing, placeholders where prices belong, and no em dash. a: FAIL T4	Set the two merge-field paragraphs in Poppins, then fill the unowned [PLACEHOLDER LINK] on-screen cards with a named owner and deadline so Philip's editor does not have to ask what URL goes there.
T3	11. Video scripts 3 lot walkthroughs + 1 build explainer	superseded	3	Ranking a Winner a: Variant a is the only entry and it is clean on brand, voice and compliance (exact palette, Poppins, no em dash, builder-pays-the-fee line in every outro, cost held as [RANGE TO BE SET BY JACOB]), but it does not pass T3. a: FAIL T3	Add a rule that maps the release lot to a script (for example: in a platted subdivision with recorded restrictions, script 2; in a platted subdivision without them, script 1; metes and bounds, script 3, using the Plat field on the parcel pull), take city and lake out of the script titles, and replace the 'if possible' and 'if allowed' shot notes with yes or no rules.
T3	11. Video scripts 3 lot walkthroughs + 1 build explainer	superseded	4	Ranking a Winner a: Variant a is the only entry and its lot walkthroughs are near operator-proof (a two-question legal-description rule picks the script, the release and parcel pull have owners and deadlines, empty fields have a cut rule), on palette, Poppins only, no em dash, and every outro carries the builder-pays-the-fee line. a: FAIL T3	Add an operator box to the build explainer that names the source, owner and release for every location and b-roll item, or a fallback rule (stock or skip) when one is not available.
T3	11. Video scripts 3 lot walkthroughs + 1 build explainer	judged	5	Ranking a Winner a: Variant a is the only entry, and it now closes the round 4 gap with a build explainer footage table that names source, owner, release and fallback for every shot, on exact palette in Poppins with no em dash and the builder-pays-the-fee line in every outro, but it still fails T3 on how the permit count is produced. a: FAIL T3	Define {{nearest_permits_1mi}} and {{nearest_permit_builders}} as a rule (new single family residence permits only, named Montgomery County or city portal, last 12 months, 1 mile radius drawn from the parcel centroid with a named map tool, builder name copied from the permit's contractor field), and fix the build explainer's 'same morning as that shoot' sunrise shot, which contradicts Script 1's 'last hour before sunset' call time.
T3	12. Builder one-pager Builder acquisition one-pager	superseded	1	Ranking a > b Winner a: Variant a hands a builder a concrete reason to sign (the lot file with flood, utilities, permits and loan type, four brief-sourced numbers, and a direct 'signing 10 builders' ask), while variant b reads as a process list that ends on a vague 'Want the terms in writing?' a: FAIL T4; b: FAIL T4	Enlarge variant a's logo to at least 1in of visible artwork with wordmark-height clear space (and add the four-color brand strip the guidelines require on builder pitch covers), then it is ready for Jacob to set the fee terms.
T3	12. Builder one-pager Builder acquisition one-pager	judged	2	Ranking a > b Winner a: Variant a makes the lot file the visible proof a builder cannot get elsewhere, states the 10-builder ask in the CTA, and covers both lot-owning and land-search buyers, while variant b hides that proof in a footnote sentence. a: pass; b: pass	Jacob still has to set the fee terms placeholder, and the lot file panel needs one approved, filled sample lot (flood zone, utilities, nearby permits) in place of the raw merge fields before a builder sees it in print; the logo is also a raster PNG where the guidelines call for SVG or PDF in print.
T4	13. Feasibility tool spec Spec	superseded	1	Ranking a Winner a: Variant a is the only entry and it is a thorough, source-cited spec with a four-state verdict, 14 explicit rules, real edge cases, compliant fee framing and acceptance tests, but it still fails the untrained-operator test. a: FAIL T3	Define the owner-name match rule exactly (normalisation, order swap, entity handling) and fix the T8 versus flood_minimal 'guarantee' contradiction; Jacob should also consider replacing 'The lot report costs you nothing' with plainer wording, since it sits close to a free claim even with the builder-pays line beside it.
T4	13. Feasibility tool spec Spec	judged	2	Ranking a Winner a: Variant a is the only entry and it is a complete, rule-based developer spec (inputs with validation, nine sourced adapters with refresh and stale windows, a 14-rule verdict engine, edge cases, acceptance tests, and approved-wording disclaimers) that never promises buildability, never invents a cost, and always states the builder pays the fee. a: pass	Jacob still has to approve the disclaimer and finding texts with the brokerage compliance contact, supply the real TREC IABS and Consumer Protection Notice URLs (currently href="#"), and decide the cost range or confirm the cost block stays hidden.
T4	14. Tracking sheet Weekly tracking sheet	superseded	1	Ranking a Winner a: Variant a is the only entry, and it gives a precise definition, owner, recorder, deadline and data source for each of 33 metrics, plus a 45 minute agenda that puts money first, but it still fails T3. a: FAIL T3	Add a CSV column or a second weekly tab for every defined metric, and add a source and persona attribution rule plus a glossary covering business hours, GHL, PIN, cron, skip trace, phase 0 and the playbook grade criteria, so a new ISA can fill the sheet without asking anyone.
T4	14. Tracking sheet Weekly tracking sheet	superseded	2	Ranking a Winner a: Variant a is the only entry and it is a thorough, on-brand, answer-first sheet with 33 defined metrics, owners, deadlines, CSVs and a timed pipeline agenda, but it still fails T3. a: FAIL T3	Add exact fill rules for hot-feed H1 to H3 rows and the TOTAL row (sum counts, recompute medians and percents from GHL), rename A7_appointments_due to A8_appointments_due, and define every named person, so an untrained ISA can fill it without asking.
T4	14. Tracking sheet Weekly tracking sheet	superseded	3	Ranking a Winner a: Variant a is the only entry and it gives every one of 33 metrics an exact definition, a named owner, a deadline and a CSV column, with no invented targets and correct fee framing. a: FAIL T3	Define HOT, WARM, COLD and DEAD with one observable rule each in the Terms table, and settle whether a hot-feed lead's non-H counts move to its persona row once known, since the persona rule and the hot-feed rule currently read as contradictory.
T4	14. Tracking sheet Weekly tracking sheet	superseded	4	Ranking a Winner a: Variant a is the only entry and it is close: every CSV column has a named filler, a source and a plain-language rule, the grades are tested in a fixed order, NA, blank and zero are defined, and the voice passes cleanly, but it still fails T3 on one unresolved operating rule. a: FAIL T3	Add one rule for weekend activity (enter Monday by 9am, or move the lock to Sunday 23:59) and say which of the six source rows M1 spend on tools and skip trace goes in, and the sheet passes T3.
T4	14. Tracking sheet Weekly tracking sheet	judged	5	Ranking a Winner a: Variant a is the only entry, and it is a near-complete operator manual: every column has one owner, one source and a deadline, it covers weekend entries, spend allocation and blank versus zero versus NA, it uses Poppins and exact brand hexes, and it has no em dash, but it still fails T3. a: FAIL T3	Add a grade, or a line in the DEAD or WARM rule, for a lead who has chosen a builder but not signed, and rename Q4 to match 'builder-signed', so every lead fits exactly one grade.