The pages never make the offer — and it's measurable
The landers are clean, purpose-built pages: no nav, prominent "Start My LLC" above the fold, Trustpilot at the top. The structure is not the problem. The message is: across all four landers there is not a single mention of price, speed, or — outside the button label — the product. A paid visitor gets ~5 seconds of identity flattery and a button, and no reason to tap it now. Meanwhile Bizee's own homepage leads with "Launch Your LLC in Minutes" and $0 packaging — the offer exists; it just isn't on the pages the ads pay to fill.
The plan, in four moves
Put the offer on the page.
Headline / subheadline / CTA-microcopy variants for each lander, written and ready below — product named in the H1, price and timing in the first screen. live audit competitorAnswer "why do they escape" with session replays, not opinions.
A 90-minute FullStory protocol below tags what logo-clickers went looking for — price, entity help, or trust — and maps each answer to a specific lander change. click-mapShip the no-risk fixes this week.
Deep-link the CTAs (verified working), fix the "Let'S File Your LLC" casing bug, triage the broken analytics beacon firing on every lander. live auditInstrument signups per lander — then the next flight optimizes to conversions, and the drop-off gets a dollar figure.
The funnel model is built and tested; it needs one input. needs data
Four clean landers with the same silence where the offer should be
The offer-word census
Occurrences in the full page text of each live lander, July 3. The product is named on the button and again in how-it-works — and that's it. Price and speed are never mentioned at all:
| Term | Side Hustle | Freelancer | Creator | Trades | Bizee homepage |
|---|---|---|---|---|---|
| "LLC" | 4 | 5 | 5 | 5 | 13 |
| "$0" | 0 | 0 | 0 | 0 | 3 |
| "free" | 0 | 0 | 0 | 0 | 1 |
| "price / pricing" | 0 | 0 | 0 | 0 | 3 |
| "minute(s)" | 0 | 0 | 0 | 0 | 1 |
Method: full rendered-page text, phone viewport, July 3. The "LLC" hits are the two CTA buttons, the how-it-works step, and the closing header. live audit
The template is good. That's what makes this cheap to fix.
All four landers share one clean structure: Trustpilot chip → H1 → subhead → CTA → segment-tailored benefits → 1M-businesses proof + three named testimonials → how-it-works → closing CTA with 148K+ ratings. Nothing needs rebuilding — the test program below only changes words in the first screen, which is also the only screen most visitors ever see (~4.7s median dwell).
A visible button nobody taps is a motivation problem
The click-map's most-cited number — "Start My LLC" ranks #3–#7 in mobile clicks, 0–1% of taps — was read as a visibility problem ("the CTA is buried"). The live pages disprove that reading: the CTA is above the fold, full-width, high-contrast, on every lander. So the number means something more useful: visitors see the ask and aren't persuaded in the ~5 seconds they give the page. That is a message problem.
What 4.7 seconds buys
One screen: headline, subhead, button. Every lander spends that screen on identity ("Your New Title? Entrepreneur") and asks for commitment ("START MY LLC") with no price, no timeline, no "what happens next." The dwell time isn't the disease — it's the budget. click-map live audit
The set's own A/B says concrete wins
The two H1s that name a concrete transformation — Freelancer's "Do the Work. Officially." (6.1%) and Creator's "Create a Business for Your Content" (5.1%) — beat the riddle (Side Hustle, 3.2%) and the generic control ("Do Business. Officially.", 3.8%). Five landers is a pattern, not proof — but the pattern and the census point the same direction. click-map
The ad → page handoff drops the story
The paid side runs UGC testimonial and talking-head video (the winning Side-Hustler ad: "Concept 4 – Testimonial," 6.96% CTR). The lander answers a real person's story with a stock photo and an abstract line. The named testimonials that could continue that story sit ~2,000px down the page. ad report live audit
The message program — offer-forward copy, per lander, ready to test
The formula, from what already works in this market and on Bizee's own homepage: H1 names the segment's moment and the product · subhead carries the offer (price + speed) · CTA microcopy de-risks the tap. Control is live today; A keeps each lander's current emotional angle and adds the product/offer; B is the offer-forward version. Control is never a strawman — it's the real test floor.
Side Hustle
3.2% CR · 164.8K PVsbiggest traffic, near-worst rate — first priority
Freelancer
6.1% CR — the winnerprotect it; iterate the subhead only first
Creator
5.1% CR — strongbrand-deal angle is its own body copy, promoted
Trades
3.0% CR, 2.52s dwellverify field performance first (below); copy second
The header CTA text
"Get Started" is the only CTA a mid-page reader can recover (the header hides on scroll-down and returns on any scroll-up), and it currently carries zero scent. Test "Start My LLC", then "Start My LLC · $0". Same slot, same design, three words of difference.
CTA microcopy — the de-risk line
One line under the hero button: "Starts at $0 + your state's filing fee. Done in minutes." ZenBusiness runs exactly this pattern; Bizee's homepage already claims "Launch Your LLC in Minutes," so the speed claim is house-approved language. One check before shipping: confirm with whoever owns pricing that leading with $0 on paid landers fits the upsell economics — it's Bizee's public homepage offer, but paid traffic is a different margin conversation.
The benchmark: how the incumbents spend their first screen (live, July 3)
| Page | H1 | Offer signals in the first screen |
|---|---|---|
| ZenBusiness /llc/ | "Start Your LLC in Minutes" | Product + timing in the H1; "Starts at $0 + state fees." as microcopy touching the CTA; Trustpilot 4.8 / 30,793 reviews directly below. The tightest product-price-timing stack in the category. competitor |
| LegalZoom LLC page | "Start your LLC with confidence" | Product in the H1; sells confidence + guarantee rather than speed; $0 held for the first package card one scroll down; named customer quote in the hero. competitor |
| Bizee homepage | "Start Your Business With Confidence" | Hero card "Launch Your LLC in Minutes" + entity/state picker; "$0 + state fee" Basic and "$199 + state fee" Standard packages below. Bizee already writes offer-forward copy — one click from the landers. live audit |
| Bizee /form/ landers | (four identity-led lines) | No product outside the button. No price. No timing. Anywhere on the page. live audit |
Both incumbents put "LLC" in the H1 of their LLC pages. They differ on the second lever — ZenBusiness sells speed, LegalZoom sells confidence — which means both positions are open to Bizee on segment landers; what neither leaves open is saying nothing.
Why do they escape to the homepage — and what were they looking for?
First, the honest scale of it: 272–402 logo-exits per lander in the data window is 7–8% of visitors who clicked anything, but ~0.3% of all visitors. It's not the leak that empties the funnel — the 96% who click nothing are that — but it's the leak that talks. Each logo-click is a visitor who wanted something the lander didn't give and went looking for it.
What the lander withholds that the homepage offers
| Visitor question | On the landers | On the homepage |
|---|---|---|
| "How much does it cost?" | Nothing | "$0 + state fee" Basic / "$199 + state fee" Standard cards |
| "How long does it take?" | Nothing | "Launch Your LLC in Minutes" |
| "Do I even need an LLC — or something else?" | Nothing | "Find the Entity That's Right For You" comparison |
| "What do I actually get?" | Three benefit blurbs | Package contents, dashboard preview, support promises |
| "Can I trust them?" | Trustpilot chip + 1M businesses + 3 quotes | Same, plus company story and FAQs |
The answer the evidence supports: they go looking for the price
Follow the paid path: the ad names no price, the lander contains no price word anywhere, and the order form reveals pricing only after two selections. The homepage is the only page in the entire journey that shows a dollar figure — and it's one logo-tap away. For a purchase where the #1 unknown is "what does this cost," the escape hatch leads exactly where the missing answer lives. (Second candidate: entity-fit — "do I even need an LLC?" — which the homepage also answers and the lander doesn't.) The protocol below confirms it with real sessions and apportions the two: if price dominates, the CTA microcopy line pays for its test before it runs; if entity-research shows up, a "Not sure an LLC is right for you?" explainer link keeps researchers on the lander.
The 90-minute FullStory protocol (runnable by anyone with view access)
Build the segment.
Sessions with a page path starting/form/AND a click on the header logo link, over the flight window (May 28 – Jun 26). Split by lander.Watch 50 sessions per lander
(or all, if fewer), starting ~5 seconds before the logo tap. Note pre-click behavior: scroll depth, taps on non-interactive content, chat opens.Tag the after-path:
PRICE (reached packages/pricing), ENTITY (entity comparison, name search, guides), TRUST (about, reviews, support), BOUNCE (left ≤10s after landing on the homepage), OTHER (note it).Decision map, agreed in advance:
PRICE ≥ 30% → ship the price microcopy + a slim package strip on the landers. ENTITY high → add the explainer link under the CTA. TRUST high → move a named testimonial (with face) into the hero region. BOUNCE high → those clicks were low-intent; deprioritize the whole question.Three special pulls while in there:
(a) the Creator desktop session with 15 hero-clicks from one user — watch what the button did; the live button is a working link today, so this is either fixed or was transient. (b) The slowest decile of Trades sessions with the page-speed overlay — adjudicates the 2.52s question below. (c) A handful of sessions with taps on non-interactive sections — those are the clicks FullStory labels "excludedUrls" (explained below); see what visitors expected to happen.
Technical findings from the live audit
A broken first-party beacon fires on every lander
Bizee's own vendor.js attempts /events?cee=no calls to two unallowlisted serverless hosts (…ecs.us-east-2.on.aws, …run.app); the site's CSP correctly blocks them — 4 console errors per page load, every lander, every visitor. Either dead code to delete or a misconfigured integration silently losing data. A LinkedIn pixel sync is also CSP-blocked. live audit
The CTAs skip a free step — deep-link them
Every lander CTA opens the bare order URL, which asks the visitor to pick an entity type — on a page they reached by tapping "Start My LLC." The order system already supports prefill: ?entityType=LLC selects LLC (verified July 3). One attribute change per lander removes one of the two steps between the tap and seeing a price. live audit
Trades' 2.52s: the lab says look at the field
On a warm cache the Trades lander is light — 582ms load, 265KB, TTFB 51ms — so raw page weight isn't the story. Two live suspects: field performance (117 requests of tag/martech JS on job-site connectivity — FullStory's speed overlay or CrUX mobile data adjudicates), and upstream coldness — Trades' ads also post the weakest CTR of the set (2.37% vs Side-Hustler's 6.96%), so the segment may be arriving least convinced. Both checks are in the plan; neither costs a rebuild. live audit ad report
Two small ones while someone's in the code
(1) The closing header on all four landers renders as "Let'S File Your LLC" — a title-case function tripping on the apostrophe. (2) The "excludedUrls" entries in the click data are not a phantom block: every content section carries an Alpine.js margin helper with an empty config, and FullStory names section-taps after that attribute. It's dead code + a labeling artifact — the real signal is that visitors tap non-interactive content. Deleting the dead directive and naming sections properly makes the next click-map readable. live audit
The plan to lock
| When | What ships | Why / what it needs |
|---|---|---|
| This weekno-risk, no-test | CTA deep-links (?entityType=LLC) · "Let'S" casing fix · vendor.js beacon triage · dead Alpine directive cleanup · run the 90-minute replay protocol |
All verified against the live pages; none change what visitors see except the casing fix. The replay protocol produces the escape answer before the copy test launches. |
| Before the next flight | Build the A/B variants per lander (copy above) · header CTA label variant · price microcopy (pending pricing sign-off) · move one named testimonial with a face into the hero on one lander as a test cell · instrument a signup event Meta can see | Words only — the template stays. The signup event is the unlock for everything in the last row. |
| During the flight | 50/50 split per lander · primary metric: order-page entries per landing-page view (measurable today) · secondary: signups per LPV once instrumented · one full week minimum per cell | At ~100K+ views per lander per flight, a meaningful lift on a low-single-digit rate is detectable inside a flight. Read weekly, decide on the pre-registered metric, not on vibes. |
| Then | Dollarize the funnel for the CFO view · shift ad sets from view-optimization to conversion-optimization | The funnel/ICE model is built and unit-tested; it needs signup counts per lander. The campaign currently tells Meta to buy views ($0.47 each) — once the conversion event flows, the same budget buys buyers. (Media call to make with whoever owns the account.) |
Four asks, to run this end-to-end
1) Signup/lead counts per lander for the flight window (unlocks the dollar view for Dorian). 2) FullStory view access (or 90 minutes of someone who has it) for the replay protocol. 3) The creative copy of the winning hooks (Concepts 4/5/12/16) to tighten ad→lander message match. 4) A yes/no from pricing on leading with "$0 + state fee" on paid landers.
The media held up its end
Worth keeping on one page for the finance view: the campaign bought reach cheaply and evenly across audiences. The waste isn't in the buying — it's in what the page does with the click.
| Audience ad set | Impressions | Landing-page views | Spend | Cost / view |
|---|---|---|---|---|
| Side Hustler | 3,340,504 | 164,394 | $56,294 | $0.34 |
| Trades | 5,167,468 | 128,829 | $56,483 | $0.44 |
| Creator | 4,049,027 | 105,941 | $56,420 | $0.53 |
| Freelancer | 4,656,210 | 92,435 | $56,447 | $0.61 |
| Retargeting (S1–S4) | 2,986,760 | 87,853 | $45,331 | $0.52 |
| Brand / Control | 2,504,634 | 64,866 | $30,076 | $0.46 |
| Total | 22,704,603 | 644,318 | $301,049 | $0.47 |
Bar widths sketch funnel shape; labelled figures are measured. Reporting window 2026-05-28 → 06-26; campaign optimized for Landing Page Views; ad sets not delivering since June 26 (flight ended) — which is exactly why the fixes above can ship before the next dollar is spent.
Methodology & sources
Claims are tagged to their source throughout. The July 3 live audit: all four /form/ landers plus the homepage and order page, rendered at a 390×844 phone viewport (Creator also at 1440 desktop), fresh US sessions; verbatim text extraction, element geometry, scroll behavior, console and performance capture; screenshots archived with the supporting docs. Competitor copy quoted verbatim from live pages the same day. The funnel/ICE math is a unit-tested model (21 cases) awaiting the signup input.
Ad-report-Meta-Growth-2.xlsx) + click-map analysis (Meta_Lander_Analysis.docx, Mar 23 – Jun 20).