NILSON
← Home

Conversion Rescue

Your app looks great.So why aren’t people converting?

Find the hidden friction turning your visitors into lost opportunities.

Paste the address of a public page. I really read it — its code, its forms, its buttons, and how it renders on a real phone through Google Lighthouse — and you get three concrete findings, each with its evidence and its fix.

Free · no account · no email · about 10 seconds, plus the mobile render

See a real report ↓

01 — What gets checked

Ten dimensions, evidence, no generic advice

Every finding states what was read or measured, its severity, my confidence, the business hypothesis, the fix, its difficulty and how to validate the change. A dimension the page does not let me judge is marked "not observed" — never "fine".

First-impression clarity

Read or measured automatically

Main headline present, vague or generic, missing subheading, leftover template text.

Needs an on-screen review (full audit)

What the eye sees first, and whether the message lands in five seconds.

Intent and offer match

Read or measured automatically

Default title and description, match between search result and page, share image, page empty without JavaScript.

Needs an on-screen review (full audit)

Whether the page answers what visitors from your channels are looking for.

Forms and smart defaults

Read or measured automatically

Number of fields, labels, field types, autocomplete, required phone, pre-ticked boxes, defaults.

Needs an on-screen review (full audit)

Error messages, the flow on a phone, what happens after sending.

Onboarding and progress

Read or measured automatically

Single-block sign-up, announced steps and duration.

Needs an on-screen review (full audit)

The real sign-up flow up to the first result.

Value before commitment

Read or measured automatically

Value given before any ask: example, demo, tool, trial.

Needs an on-screen review (full audit)

Whether that value is really useful to your audience.

Personalisation and user-created value

Read or measured automatically

Several audiences without a dedicated path, no personal result.

Needs an on-screen review (full audit)

Which audience really matters, and what they want to produce.

Perceived risk and real losses

Read or measured automatically

Unverifiable urgency, missing cancellation or refund terms, email asked without explanation.

Needs an on-screen review (full audit)

Your customers’ real fears before paying.

Pricing and choice architecture

Read or measured automatically

Prices nowhere, crossed-out prices, number of plans, recommended plan.

Needs an on-screen review (full audit)

Whether comparing the plans really helps choosing.

Calls to action and decision paths

Read or measured automatically

Main action missing or far from the headline, generic labels, competing actions, dead or test-address links.

Needs an on-screen review (full audit)

The visual hierarchy of the buttons and the decision path.

Trust, accessibility, mobile, performance

Read or measured automatically

HTTPS, mobile, zoom, language, alt text, legal mentions, proof, footer year, builder badge, speed, stability, accessibility and JavaScript errors (Lighthouse).

Needs an on-screen review (full audit)

How credible the proof is, the quality of visuals, the tone.

How the analysis works

The page code
I download the public page like a browser, from my server, identifying myself. Headlines, buttons, forms, prices, links, legal mentions: everything is read in the code received. The HTML is not kept.
A real mobile browser
Google PageSpeed Insights opens the page in Chrome on a simulated phone: speed, stability, accessibility, JavaScript errors, and a capture of the first screen. Pages built in JavaScript are visible there.
Written rules
Every finding comes from an explicit rule with its threshold (WCAG, Core Web Vitals, consumer law). No randomly generated text: the same page gives the same report.
An explained score
The score adds up the friction observed, weighted by severity, my confidence and your goal. The formula is printed under the score.

This analysis is an assessment, not a measurement: it knows neither your traffic nor your conversion rate, and no revenue figure is derived from it. It is not a guarantee of results.

Example

The real report, on a demonstration page

Rather than exposing someone’s site, I wrote a fictional page that looks like a first AI-generated version. The analyser reads it live, with exactly the same rules as for your page: what follows is its real output.

Fictional demonstration page (Orvane does not exist), written to look like a first AI-generated version. The report below is computed live by the analyser, with the same rules as for your page. See the fictional page ↗

Opportunity score

100 / 100

High margin

Lots of fixable friction, including strong points. Start with the three below.

Coverage: 100% of dimensions observed

By dimension

  • First-impression clarity25.9 pts
  • Intent and offer match10 pts
  • Forms and smart defaults45.9 pts
  • Onboarding and progress1.4 pts
  • Value before commitment0 pts
  • Personalisation and user-created value0 pts
  • Perceived risk and real losses11.9 pts
  • Pricing and choice architecture7.7 pts
  • Calls to action and decision paths33.7 pts
  • Trust, accessibility, mobile, performance24 pts
How this score is calculated

Every finding adds points: severity (high 15 · medium 7 · low 3) × confidence (high 1 · medium 0.7 · low 0.4) × the weight of its dimension for your page type and goal. Capped at 100. Dimensions not observed do not count: with partial coverage the score is a floor. It is not a conversion rate, and no revenue figure is derived from it.

Calls to action and decision paths

Severity
High
Confidence
High
Difficulty
Quick (under an hour)
Source
Read in the page code
Element
href / src / action

What was read

  • localhost

Why it matters

Links, images or a form target a development address (localhost, a preview deployment…). For a visitor those links are broken, and a form may send data to the wrong place.

Hypothesis

Fixing these addresses may prevent broken paths and lost requests.

Recommended fix

Replace them with production addresses (ideally relative paths or an environment variable) and re-test every form.

How to validate — Manual check

Check it yourself on the live page, on desktop and on a phone.

First-impression clarity

Template text was left in the page

Severity
High
Confidence
High
Difficulty
Quick (under an hour)
Source
Read in the page code
Element
texte visible / visible text

What was read

  • « Lorem ipsum »
  • « dolor sit amet »
  • « consectetur adipiscing »
  • « John Doe »
  • « Acme Inc »

Why it matters

Placeholder text ("lorem ipsum", "Your Company", "John Doe"…) is visible. Nothing signals an unfinished site faster, and it casts doubt on everything else.

Hypothesis

Removing these template leftovers may prevent immediate exits among the most attentive visitors.

Recommended fix

Search the project for these words, replace them with real content or remove the section. Re-read the whole page on a phone.

How to validate — Manual check

Check it yourself on the live page, on desktop and on a phone.

Forms and smart defaults

Severity
High
Confidence
High
Difficulty
Quick (under an hour)
Source
Read in the page code
Element
form 1 (lead)

What was read

  • « Send me news and special offers »

Why it matters

A newsletter, marketing or consent box is pre-ticked. In the EU a pre-ticked box is not valid consent (GDPR recital 32; CJEU Planet49, 2019), and visitors who notice it lose trust.

Hypothesis

An unticked box by default protects trust and the real value of your email list.

Recommended fix

Untick it by default. A good default is reversible and never decides a purchase or a consent on someone’s behalf.

How to validate — Manual check

Check it yourself on the live page, on desktop and on a phone.

Lever · Smart defaults A reasonable, visible, reversible default where the context supports it. Never a pre-selected purchase, consent or sensitive choice.

Run the same analysis on my page

02 — Psychology, with guardrails

Six behavioural levers — used honestly, or flagged

The analysis does not tick a fixed list: it only applies a lever where the page gives it a hold, and it flags pages that turn these levers against their visitors. This page applies them too, under the same rules.

Smart defaults
Here: your conversion goal is pre-filled from your page, and editable. Never a pre-ticked purchase or consent — the audit flags pre-ticked boxes.
Goal gradient
Here: each step of the analysis shows when it is really done, and your plan counts the decisions you actually made.
Reciprocity
Here: three complete findings before any request for an email or a payment. No hidden catch.
IKEA effect
Here: you build your own action plan — fix, learn, delegate — before being offered to save it.
Loss aversion
Here: documented consequences (WCAG thresholds, Core Web Vitals, law), never an invented lost-revenue figure or urgency. The audit flags unverifiable countdowns.
Contrast
Here: the three options are compared on their real scope and real starting price. The audit flags crossed-out prices to check.

Guardrail

03 — Three ways to continue

Do it yourself, learn, or delegate

All three options are real and compared on what they contain. Start with the free analysis: it is often enough to fix the essentials yourself.

Free analysis

To know where it hurts, now.

Free

Included

  • Automated analysis of one public page
  • Three complete findings, with evidence and fix
  • Explained opportunity score
  • Your action plan (fix · learn · delegate)
  • Private link, saved by email

Not included

  • No human review
  • One page, not a full flow
  • Flows behind a login are not seen
Scan my page

Conversion Audit

For a prioritised roadmap, checked by a person.

from CHF 490

Report within 5 business days of payment · Standard price once the scope is confirmed, accepted in writing before payment.

Included

  • I go through your key pages and one flow on desktop and phone
  • All ten dimensions, finding by finding, with evidence and real screenshots
  • Hypothesis, fix, difficulty and validation method for each
  • Roadmap: now · next · later
  • Downloadable report (PDF through print)
  • One clarification exchange after delivery

Not included

  • No change to your site
  • No conversion-rate promise
  • No A/B test run for you
Request the full audit

Conversion Fix Sprint

To get the fixes done, without spending your evenings on them.

from CHF 1’490

Scope and timeline set in the quote · Always a written quote, accepted before any payment.

Included

  • The chosen fixes, implemented (code, CMS or design files depending on your site)
  • Delivered as a pull request or documented changes
  • Measurement of key steps set up if needed
  • One rework round included, defects fixed for 14 days

Not included

  • Nothing is published without your written approval
  • Not a full redesign (see Design Rescue)
  • No promised figures
Request a quote
Do it yourself, learn, or delegate
FreeAuditSprint
Who does the workThe automated analysisMe, on screenMe, in your project
Pages coveredOne public pageKey pages + one flowPer the quote
EvidenceCode + Lighthouse renderPlus human review and screenshotsDocumented before / after
Changes to your siteNoneNoneYes, with your approval
Your timeA few minutesA brief, a readA brief, an approval
  • "From" prices are starting points: the final price is the written quote you accept.
  • Online payment through Stripe, after accepting the quote. Nothing is charged before.
  • I work alone: few projects in parallel, a reply within 24 hours Monday to Friday.

Request

Request the full audit or a quote

A few lines are enough. I read them myself and reply within 24 hours, Monday to Friday: a question, a written proposal, or a reasoned no. Nothing is charged before you accept a quote.

What you are asking for
A public page. I never ask for access or a password.
Under 10,000 visits a month an A/B test rarely concludes: I then propose usability tests.
For example: home → pricing → sign-up.
Do not paste keys or passwords: they would be removed.

You receive a link to your project space. Nothing is charged at this stage.

Already sent a request? Find my project space

05 — The right service

It may not be a conversion problem

Frequently asked questions

Is the free analysis really free?
Yes: no account, no email, no card. The email is only used to send you the report link if you want to keep it. It is limited to a few scans per day and per person, to protect the sites being scanned.
Does an AI write the report?
No. Every finding comes from a rule written in advance, applied to what was read in the page code or measured by Google Lighthouse. The same page gives the same report, and every finding shows its evidence.
Why does my website look good but not convert?
Most often: a headline that does not say what you do, a main action lost among others, a form that is too long, no proof, prices nowhere to be found, or a slow first screen on phones. The analysis checks these on your page, and the guide explains how to fix them yourself.
My site was generated with Lovable, Bolt or v0. Does it work?
Yes. The analysis spots what these tools often leave behind: default title, links to a preview address, the builder’s badge, template text, a page that is empty without JavaScript. The Google Lighthouse render also reads pages built in JavaScript. See the guide on AI-generated sites.
Is the score my conversion rate?
No. It is an assessment of the friction observed, calculated by a formula printed under the score. Your conversion rate is measured in your analytics tool — the full audit helps you measure it properly.
What does the full audit contain?
I go through your key pages and one flow on desktop and phone, and write a report on the ten dimensions: evidence, severity, confidence, hypothesis, fix, difficulty and validation method for every finding, then a prioritised roadmap. From CHF 490, on a written quote accepted before payment.
What if I want you to make the fixes?
That is the Conversion Fix Sprint, from CHF 1’490: I price the chosen fixes in a written quote, implement them, and nothing is published without your approval.
Do you promise more conversions?
No, and be wary of anyone who does. I promise what I can keep: verified findings, fixes delivered as agreed, and a way to measure whether they work.
What happens to my data?
Your page HTML is not kept, nor is your IP address (only an encrypted fingerprint for quotas). The page address is sent to Google PageSpeed Insights for the mobile render. An unsaved report is deleted after 30 days, a saved report after 180 days. Details in privacy.

Three concrete findings, in under a minute.

Paste your page address. If the analysis is enough, fix it yourself. If not, you will know exactly what to ask for.

Scan my pageRead the guide: a good-looking site that does not convert →

A question first? danilson.daveigaramos.pro@gmail.com