Inbound auditHome, pricing, /demo, vs Apollo, a customer story, /dialer
Every trial button opens a demo form. Fix the page before buying more clicks.
A paid click lands on a page with no proof in the first screen, asks for a free trial and gets a two-field demo form instead. Five changes, none of them new spend, decide what every later dollar of paid buys.
0self-serve sign-up paths; /demo’s “Start free trial” schema points to a 404
7 vs 14trial days: pricing page vs the vs-Apollo post
0.25layout shift on /demo for real users; good is under 0.1
18,231pxto the first in-article CTA in the vs-Apollo post
01The path
From the ad to a sales call, in five steps.
“Leak” marks a step where a buyer with intent can leave. Three of five.
- Ad or searchGoogle Ads with enhanced conversions. No Meta or Reddit pixel, so only Google can build an audience from a visit.
- HomeNo proof in the first screen. Customer logos start at 1,631px, G2 stars at 12,545px.Leak
- “Get free trial”Every trial button on the site goes to /demo. /signup returns 404.Leak
- /demoTwo fields: company email and size of the sales team. No role, no current tool.Leak
- RouterDefault.com decides who gets a call. What a 1–2 seat lead sees next can't be seen without submitting.
02Page by page
The pages paid traffic lands on.
Six pages a paid click or an AI answer can send a buyer to. The dialer page already does what the others should.
| Page | What's there | The leak | Fix |
|---|---|---|---|
| Home | “Agents build your pipeline. Your team closes it.” One CTA: Get free trial. | A visitor from an ad sees a product mock and no proof. | Put G2 rating and three logos under the headline. |
| Pricing | Startup $600 a month for 2 users. Growth and Elite: custom. | Startup offers a 7-day trial. The vs-Apollo post promises a 14-day free trial. | One trial length, everywhere. |
| /demo | “See Amplemarket in action.” Submit button says “Get free trial”. | Field layout shift 0.25, 2.5× Google's “good” line of 0.1. Lab, slow 4G: the form renders after 12.4s; real users p75 1.4s. | Reserve the form's space; load the router after the first paint. |
| vs Apollo | 4,190 words, FAQ schema, ranks first on Google. | First call to action in the body at 18,231px. | A CTA after the verdict table, with the email field prefilled into /demo?email=. |
| O'Reilly story | “$21M in 90 days.” Your strongest number. | No call to action in the body. “Watch Demo” links to #. | End the story with the trial form and a 90-day plan offer. |
| /dialer | Inline email and G2 stars in the hero. | None found. This is the pattern to copy. | Use it as the template for the pages above. |
03What a demo may cost
At 12-month payback, a booked demo is worth up to $864.
The price is public. The other three numbers are my assumptions, marked in grey; on day 1 they come out of HubSpot. Growth and Elite deals raise the ceiling.
$7,200Revenue in year one, Startup plan$600 a month × 12, public price
$5,760Most you can pay to win one customer and earn it back in 12 months× 80% gross margin (assumed)
$1,152Per demo that happens× 20% of held demos win (assumed)
$864Per demo booked: the number paid is judged on× 75% of booked demos are held (assumed)
The fixes below move the last two numbers. Lift the show rate from 75% to 85% with a role field that routes founders to a founder-led call, and every booked demo is worth $979 instead of $864: more room to bid on “apollo alternative” with the same payback.
04Five fixes
In order of cost: copy, form, code.
Each one has the number that says whether it worked. None needs new budget.
- One trial, one lengthPick 7 or 14 days and use it on pricing, blog and ads. A buyer who reads 14 and gets 7 starts the sales call doubting you.Measure: Count of trial requests that ask about length; should go to zero.
- Proof in the first screenG2 rating and three logos under the home headline: the dialer page already does it.Measure: A/B on paid traffic only: demo requests per visit.
- A CTA where reading endsThe vs-Apollo page asks for nothing in the article until 18,231px. Add the form after the verdict table and prefill the email into /demo.Measure: Demo requests from the page, per 1,000 sessions.
- Two more fields on /demoRole and current tool. Current tool routes Apollo switchers to the vs-Apollo story; role keeps founders off an SDR call.Measure: Form completion rate before and after. If it drops more than 10%, make current tool optional.
- Stop the form from jumpingField layout shift on /demo is 0.25. Reserve the form's height and load the router after first paint.Measure: CrUX CLS under 0.1 within 28 days.
05Tracking
Google can learn from a lead. Nobody else can.
Read from the page's tags and network calls with the cookie banner left untouched.
On the site
- Google Tag Manager and GA4, consent mode on
- Google Ads with enhanced conversions and a lead event
- HubSpot forms and analytics
- Default.com form and router
- PartnerStack for partners; Hotjar behind consent
Not found
- Meta or Reddit pixel: no audience for retargeting outside Google
- Website visitor identification (RB2B, Clearbit, 6sense): none fired with consent denied; worth checking after consent
- G2 buyer-intent tag
- Opportunity sent back to Google Ads: not visible from outside; it's the first thing I'd check
06Method
How this was checked.
The raw audit is published.
- Chrome DevTools in a clean browser profile, from a Portuguese IP, cookie banner not accepted. No form was submitted, so what happens after the router is unseen.
- Speed: lab traces at 390px on slow 4G with 4× CPU slowdown, next to Google's field data (CrUX, 75th percentile) where it exists. The lab overstates; the field number is the one to fix against.
- Pixel positions are measured at 1440px wide.