Why Your Website Isn’t Converting: The WebCraft Conversion Diagnostic
“My website doesn’t convert” describes at least eight different problems, and they don’t all deserve the same attention. This diagnostic helps you find which one you actually have — and what to fix first.
“My website doesn’t convert” is a single sentence describing at least eight different problems. It might mean visitors can’t tell what you do. It might mean they understand you fine but don’t want what you’re selling at the price you’re implying. It might mean they want it, believe you, and then hit a fifteen-field form on a phone at a job site.
Those are different failures with different fixes, and the expensive mistake is treating them as one thing. Most conversion advice hands you a list of improvements and lets you pick whichever looks easiest — which is how businesses end up with a faster, prettier version of a website nobody understood in the first place.
This guide does something narrower: you’ll score your website across eight areas, find the ones genuinely broken, and work through them in dependency order.
That order is the whole argument here. Problems upstream change what problems downstream even mean. If a visitor can’t tell what you sell, your call-to-action wording is irrelevant — they were never going to click it. If your offer isn’t compelling, better proof just makes a case for something people don’t want. Fixing downstream symptoms while upstream blockers remain is the most common way to spend real money on a website and see nothing change.
The journey behind that order is simple: someone arrives, works out what you do and whether it applies to them, decides whether they want the outcome, decides whether they believe you can deliver it, and then either acts or doesn’t. Each stage depends on the one before it.
Who this is for
This is written for owners and operators of businesses that sell something considered — services, trades, professional practices, custom work, B2B products — where someone researches before they contact you.
It’s less useful for high-volume ecommerce, where the diagnostic questions genuinely differ, or if you’re pre-launch with no live site to score. With no website yet, the project planner is a better starting point. If you’re selling products and still choosing a setup, Commerce 101 covers that ground first.
What this guide cannot tell you
This diagnostic examines your website — and the most common cause of “the website doesn’t work” isn’t the website. It cannot tell you whether the market wants what you’re selling, whether your pricing is viable, or whether you have enough demand to sustain the business. A site only converts intent that already exists: if nobody wants the offer, a perfect site turns a small number of people into a small number of sales, and the problem was never the site. Industry, season, and local competition sit outside its reach too. Keep that in view as you score.
By the end you should be able to answer three questions:
- Why is my website underperforming?
- What should I fix first?
- Do I need a small repair, a larger redesign, or a complete rebuild?
How to Use This Guide
One rule matters more than the rest, and it's the one people break immediately.
Score what a visitor actually experiences, not what you intended to build. You cannot un-know your own business while reading your own homepage. You know the phone number is in the footer and the third menu item leads to pricing. A visitor knows none of that, and the diagnostic is only worth doing from their position. Four rules make that practical.
The evidence rule
You may not score a category above 3 unless you can point to specific evidence on a specific page.
Not “we have testimonials” — which testimonials, on which page, next to which claim. Not “the site is fast” — fast measured how, on what connection. If you can’t name the thing, the score is 3 or below. This rule is what separates a diagnostic from a flattering self-assessment, and honest scores usually come in well below first instinct.
The blocker rule
Any scored category at 2 or below is a blocker, and a blocker outranks your total score.
A blocker doesn’t get averaged away. If Message Clarity comes in at 1, that is your finding, regardless of what the other seven categories say.
The ceiling rule
A strong overall score does not erase an upstream blocker.
This isn’t a claim that your conversion rate is mathematically capped at some number. It’s an observation about sequence: a severe Clarity or Offer problem tends to make downstream improvements largely irrelevant until it’s resolved, because visitors are leaving before they reach the part you improved. A site scoring 34 with Clarity at 2 usually behaves more like its Clarity score than its total.
The test protocol
Score the site:
- on a real phone, not a desktop browser resized narrow
- preferably on cellular data rather than your office Wi-Fi
- in a private or incognito window, so you get an uncached first visit
- reading as if you’d never heard of the business
A desktop browser with a warm cache and an owner who wrote the copy is the single most misleading way to evaluate a website.
The disambiguation rule
Several categories sit next to each other, and it’s easy to punish the same defect three times and land in a lower band than you earned. Count each defect once, in the category that owns it:
| If the problem is… | Score it under |
|---|---|
| Whether they understand you, want it, or believe you | Clarity, Offer, Trust — in that order |
| The button’s wording, or a missing next step | CTA and Conversion Path |
| The number of steps or fields once they’ve decided | Friction |
| Fine on desktop, broken on phone | Mobile |
| Slow to load or respond anywhere | Speed |
The scale
Each category is scored 0 to 5 against observable conditions, set out in the scoring block at the end of every chapter. Eight categories, five points each — 40 points total. Traffic is deliberately not one of them; that’s the gate you go through first.
GATE 00 · Unscored — do this before anything else
Is this a conversion problem at all?
Conversion problems often start before anyone lands on the page. If the people arriving aren’t looking for what you sell, a well-built site will still struggle, and scoring it tells you nothing useful.
A roofing company attracting visitors who searched for DIY roof repair tips can have excellent traffic numbers and terrible lead generation. Nothing is wrong with the website — those visitors are researching how to avoid hiring a roofer. Likewise, a software company ranking for broad informational searches may pull in thousands of readers with no intention of buying software, and no page-level fix changes that.
So the first question isn’t how much traffic do I have? It’s why are these people here?
Enough traffic to see a pattern
Before you can read conversion results, you need enough of them to read. These are working diagnostic ranges we use to decide how much confidence to place in what we're seeing — they are not universal statistical thresholds, and nothing here should be treated as a significance test.
- Under roughly 100 qualified visits. Usually too little evidence to draw strong conclusions. Obvious breakage is still worth fixing, but percentages will swing wildly on a handful of visits.
- Roughly 100–300 qualified visits. Clear problems often become visible. Treat any percentage as directional rather than precise.
- Roughly 300+ qualified visits. Patterns become more useful — but only if the traffic sources and visitor intent are reasonably comparable across the period you’re looking at.
The word doing the work in all three is qualified — a thousand visits to a blog post unrelated to your service isn’t a thousand visits for these purposes.
Conversion events matter at least as much as sessions. Five meaningful inquiries give you far less to reason from than fifty, however many sessions sit behind them. At single-digit conversions you're looking at anecdotes: fix what's obviously broken and don't over-interpret the numbers.
Intent questions
Work through these against your analytics. A “no” on several of them — particularly high blog traffic alongside flat service-page traffic — points to a mismatch rather than a website problem.
- ✓Which pages receive the most traffic, and what search or referral likely brought someone there?
- ✓Does each of those pages match the intent behind that visit?
- ✓Is the typical visitor to those pages someone who could plausibly buy from you?
- ✓Are your rankings for commercial searches, or mostly informational ones?
- ✓Do your highest-traffic pages connect to a service, offer, or conversion path at all?
- ✓Does social traffic produce any meaningful action, or only sessions?
- ✓Is the traffic inside the geography you actually serve?
Traffic without intent can make analytics look healthier than the business actually is.
Your gate outcome
A. Not enough qualified traffic to diagnose reliably. Fix anything visibly broken, then put your effort into acquisition. Scoring a site almost nobody visits produces confident conclusions built on noise — come back when you have enough visits and conversion events to see a pattern.
B. Enough traffic, but the intent is wrong. Your problem is upstream of the website. Work on what you rank and advertise for, and where that traffic lands. Return once the arriving audience could plausibly buy.
C. Enough relevant traffic to proceed. Continue to the eight scored categories.
CHAPTER 01 · Tier 1 — Can they understand you?
Message Clarity
Clarity is comprehension, and only comprehension. Within a few seconds of landing, a visitor should be able to answer three things: what does this company do, is it for someone like me, and what do I do next. When those answers are unclear, everything underneath the hero becomes harder to sell, because the reader is still working out whether they're in the right place.
Whether the offer is appealing is a separate question, scored in the next chapter. Keep them apart or you'll score the same problem twice.
The five-second test
Show your homepage to someone unfamiliar with the business for five seconds. Then hide it and ask them what the company does, who they think it serves, and what they'd click if they wanted to buy. If the answers come back vague, your positioning probably is too — and the useful part is that they'll usually tell you which of the three failed.
What weak clarity looks like
Weak homepage headlines tend to describe an attitude rather than a service. They read as polished and communicate almost nothing, because every word in them could belong to any company in any industry.
A stronger structure is outcome plus audience plus specificity: what result, for whom, in what circumstances.
Says nothing
- Innovative Solutions for a Better Tomorrow
- Transforming Ideas Into Possibilities
- Building Digital Experiences That Matter
- Your Trusted Partner in Excellence
- Full-Service Solutions for Modern Business
Says something
- Emergency AC repair in Tucson — same-day service, seven days a week
- Bookkeeping for restaurant owners who are three months behind
- Physical therapy for runners recovering from knee injuries
- Commercial electrical work for property managers in Denver
- Estate planning for families with a small business to pass on
Every headline on the right survives the same test: a stranger can tell within a moment whether it applies to them. None of them are clever, and that’s deliberate — cleverness costs a beat of comprehension, and you don’t have a beat to spare.
Diagnostic questions
- ✓Can a stranger name what you sell after five seconds on the homepage?
- ✓Can they tell whether it applies to someone in their situation?
- ✓Is the primary next step obvious without scrolling?
- ✓Does the headline describe an outcome rather than an attitude?
- ✓Is the audience named or clearly implied?
- ✓Does your navigation use words a customer would use, not internal labels?
- ✓On a phone, are all three answers available above the fold?
Score yourself
0–5
Message Clarity
- Score 0
- A stranger cannot tell what you sell from the homepage. The headline describes values, ambition, or process rather than a service.
- Score 2
- They can tell what you sell, but not whether it applies to them. No audience is named, and the next step requires hunting through navigation.
- Score 4
- What you do, who it's for, and the next step are all answerable in about five seconds — but at least one requires scrolling or a second look.
- Score 5
- All three are answerable above the fold, on a phone, by someone who has never heard of your business.
Scores of 1 and 3 sit between the anchors above.
Evidence required
Name the headline and subhead a stranger would read, and the audience they identify. If you cannot quote them from memory, score 3 or below.
Message Clarity: ___ / 5
Fix first: the headline, then the subhead, then the primary button label — in that order, and before anything else in this guide. Clarity is the cheapest fix available and the one that changes what every later fix is worth.
CHAPTER 02 · Tier 1 — Can they understand you?
Offer Strength
A perfectly clear website can still fail, because being understood and being wanted are different things. Offer Strength covers desirability and terms: is what you're asking worth doing, and is the ask sized correctly for where the visitor is?
An offer is not simply what you sell. It's the combination of six things, and weakness in any one of them can stall an otherwise healthy site:
- The problem you're addressing, stated in the customer's terms
- The desired outcome — what's different afterward
- The solution — how you get them there
- The perceived value — why the outcome is worth the price
- The perceived risk — what they're afraid will go wrong
- The next step — what you're actually asking them to do right now
Most weak offers aren't weak because the service is bad, but because only two or three of those six are ever stated and the visitor is left to assemble the rest.
The same service, framed two ways
Service, described
- Landscaping Services
- Family Law Attorney
- Business Consulting Services
- IT Support and Managed Services
- Custom Cabinetry
Outcome, offered
- Yard maintenance that keeps your property inside HOA standards — flat monthly rate, no per-visit surprises
- Divorce representation for parents who want to stay out of court, with flat-fee mediation packages
- We fix month-end close for finance teams stuck in spreadsheets — most clients close in days, not weeks
- Same-day on-site IT support for dental practices, with a named technician who knows your setup
- Built-in storage for older homes with walls that aren't square — measured, built, and installed by the same person
The left column names a category; the right names a change in someone's situation, and who it's for. The second framing is harder to write because it requires knowing precisely who you serve — which is exactly why most sites default to the first.
Reducing perceived risk
Every considered purchase carries risk, and unanswered risk stalls otherwise-interested buyers. Identify the specific worry and put its answer on the page — vague reassurance doesn't work where a specific answer does.
| What they're thinking | What reduces it |
|---|---|
| “How much is this going to cost?” | Starting ranges, a pricing framework, or what drives the number |
| “What happens after I contact you?” | The next three steps, including who calls and when |
| “How long will this take?” | Typical timelines, and the variables that move them |
| “What if it goes wrong?” | Warranty, guarantee, or how you handle problems |
| “Am I getting trapped in a sales call?” | What the call is, what it isn’t, and how long it takes |
| “Is this for a business my size?” | The size and type of client you serve — including who you don’t |
“Can I trust them?” is deliberately absent: that's Trust's question, not Offer's.
Diagnostic questions
- ✓Does the site state the problem you solve in the customer's words?
- ✓Is the outcome described, not just the service category?
- ✓Is it clear who the offer is designed for — and who it isn't?
- ✓Does the site explain what makes your approach different in a way a competitor couldn't copy verbatim?
- ✓Is there any pricing signal at all, even a range or a framework?
- ✓Is the process explained before someone has to commit?
- ✓Does the page say what happens after they contact you?
- ✓Is the size of the ask proportionate to how far along the visitor is?
Score yourself
0–5
Offer Strength
- Score 0
- The site lists services with no outcome, no audience, and no pricing signal. A visitor cannot tell what changes for them or what it costs.
- Score 2
- The service is described and the outcome is implied, but pricing, process, and next steps are absent. Common objections go unanswered on the page.
- Score 4
- Outcome, audience, and process are all stated, and there is some pricing signal. One or two significant objections are still unaddressed.
- Score 5
- All six parts of the offer are present on the page, and the most common buyer objections are answered before the visitor has to ask.
Scores of 1 and 3 sit between the anchors above.
Evidence required
Name the page where a visitor learns what it costs and what happens after they contact you. If neither exists on the site, score 2 or below.
Offer Strength: ___ / 5
Fix first: publish a pricing signal and describe what happens after contact. Those two answer the most common unspoken objections, and neither requires redesigning anything.
CHAPTER 03 · Tier 2 — Will they believe you?
Trust and Proof
People don't act when uncertainty outweighs desire, and Trust is the category that closes that gap — not by claiming competence, but by showing evidence of it.
Trust is not created by writing we pride ourselves on quality and customer satisfaction. Anyone can write that sentence, including businesses that shouldn't, and a claim every competitor can make carries no information. Evidence does.
Not all proof is equal
Most conversion advice hands you a list of fourteen trust signals and implies they're interchangeable. They aren't — proof gets stronger as it becomes more specific and more verifiable by someone who doesn't know you.
Stronger proof is specific, attributable, and checkable: named testimonials describing a particular situation and result, case studies with the starting problem and what changed, before-and-after documentation of real work, third-party reviews on platforms you don't control, and legitimately documented outcomes where you have them.
Weaker proof is real but easy for anyone to assert: years in business without context, anonymous or initials-only testimonials, association badges and logo rows, stock photography, and generic claims about quality or service.
Weaker signals aren't worthless — a missing license is a genuine problem, and certifications matter in regulated trades. They simply carry little weight alone, because a competitor can claim the same thing in the same words. The test isn't whether you have proof; it's whether a stranger could check any of it. If your entire trust story sits in the weaker group, you have a Trust problem however much of it there is.
What the difference looks like
Weak proof
- Great service! Highly recommend. — J.S.
- Over 20 years of combined experience
- We pride ourselves on quality workmanship
- Trusted by businesses nationwide
- A gallery of unlabeled project photos
Strong proof
- They replaced our rooftop unit in one day during our busiest week — we never closed. — owner, family restaurant
- Licensed and bonded, with the license number published on the site
- Every job includes a written scope before work begins, and we honor the quote
- Forty-one reviews averaging 4.8 on Google, linked directly
- Project photos captioned with the problem, the work done, and the timeline
Those illustrate form, not real WebCraft clients or results. Strong proof names a situation, a constraint, and an outcome. Weak proof names a feeling.
Put proof near the decision
This is the highest-value change most sites can make in this category, and it costs nothing but rearrangement.
The common mistake is collecting every testimonial onto a dedicated testimonials page. That page is a warehouse: it asks the visitor to interrupt what they're doing, go looking for reassurance, and come back. Hesitant people don't go looking. They leave.
Proof works where the doubt occurs — and doubt is predictable, so placement can be too:
Homepage. The doubt is is this a real business that serves people like me? Answer it with your strongest single piece of evidence near the top — a specific named testimonial, a linked review count, a genuine result — rather than a row of logos.
Service pages. The doubt is have you done this exact thing before? Answer it with work from that specific service, not general praise. A roofing testimonial on the gutter page is worth a fraction of a gutter testimonial.
Pricing or cost pages. The doubt is is this worth it, and will the number move? Value evidence belongs here: what the price included last time, what a comparable project cost, whether the quote held. This is the most under-used proof placement on almost every site we look at.
Contact page and forms. The doubt is what am I signing up for? One sentence beside the submit button — who responds, how fast, what the call is and isn't — does more work here than a page of testimonials elsewhere.
Next to any strong claim. If you say you're fast, put the evidence beside it. An unsupported claim doesn't sit neutrally; it invites the reader to start discounting everything else on the page.
Diagnostic questions
- ✓Does any proof appear on the homepage above the point where most visitors stop scrolling?
- ✓Does each service page carry proof specific to that service?
- ✓Is there value evidence anywhere near pricing?
- ✓Does the contact page say who responds, how fast, and what happens on the call?
- ✓Are testimonials attributed specifically enough that a stranger would find them credible?
- ✓Does any strong claim on the site sit without evidence beside it?
- ✓Are there real photographs of your work or team, rather than stock imagery?
- ✓Is contact information visible without hunting — including a physical location if you have one?
Score yourself
0–5
Trust and Proof
- Score 0
- No proof of any kind on the site, or nothing beyond stock photography and generic claims about quality.
- Score 2
- Proof exists but is quarantined — a testimonials page or a logo row — and never appears next to a claim, a price, or a decision point.
- Score 4
- Specific, attributable proof appears on service pages near the claims it supports. Pricing or contact points are still missing their evidence.
- Score 5
- Proof sits at every decision point — homepage, service, pricing, and contact — and is specific enough that a stranger could verify it.
Scores of 1 and 3 sit between the anchors above.
Evidence required
Name the piece of proof a visitor encounters immediately before your primary conversion point. If there isn't one, score 2 or below.
Trust and Proof: ___ / 5
Fix first: move your single best piece of existing proof next to your primary call to action. You almost certainly already have the asset; it's in the wrong place.
CHAPTER 04 · Tier 3 — Can they act?
CTA and Conversion Path
This chapter owns three questions: does a next step exist, does the page lead there naturally, and does the call to action explain what happens. What it costs to complete that step belongs to the next chapter.
A visitor should rarely reach the end of an important commercial page and wonder what to do now. Every page that could contribute to revenue needs a purpose, and its call to action should match that purpose rather than defaulting to whatever the template shipped with.
Name the outcome, not the mechanic
Weak CTA copy describes what the button does technically. Strong CTA copy describes what the visitor gets — a difference that costs nothing and pays off at exactly the moment of decision.
Describes the mechanic
- Learn More
- Submit
- Contact Us
- Get Started
- Click Here
Describes the outcome
- See Website Pricing
- Request My Estimate
- Talk About Your Project
- Book a 15-Minute Call
- See Recent Projects
Add one line under the button naming what happens next — we reply within one business day, or no obligation, and we'll tell you if we're not the right fit — and you've answered the question that was going to stop them.
Match the ask to buyer readiness
Not every visitor is ready for the same commitment, and a site offering only one door loses everyone standing at a different one. Someone early in the process may want a pricing guide, a comparison, or examples of past work; someone close to buying wants a quote, a consultation, or a proposal. A good site supports both without turning every page into a wall of competing buttons — typically one primary action, with one lower-commitment option beside it.
The mismatch to watch for is a site whose only call to action is high-commitment. If “book a consultation” is the sole path, everyone still researching has nowhere to go but away.
The path, not just the button
A call to action can be well-written and still fail because the page never built toward it. Walk a service page from the top and ask whether each section moves someone closer to acting — relevance, then desirability, then credibility, then the ask. Pages that ask before establishing credibility convert poorly no matter how good the button copy is.
Diagnostic questions
- ✓Does every commercial page have one obvious primary action?
- ✓Does the button text name the outcome rather than the mechanic?
- ✓Is there a line explaining what happens after they act?
- ✓Is there a lower-commitment option for visitors who aren't ready to buy?
- ✓Does the page sequence build toward the ask rather than leading with it?
- ✓Is the primary action available without scrolling to the very bottom?
- ✓Do your service pages tell the visitor how to get started at all?
Score yourself
0–5
CTA and Conversion Path
- Score 0
- Important pages end without any next step, or the only path is a contact link in the navigation.
- Score 2
- A call to action exists but uses generic wording, appears only at the bottom, or competes with several equally-weighted options on the same page.
- Score 4
- Each commercial page has one clear primary action with specific wording. What happens next is unstated, or there is no option for visitors who aren't ready.
- Score 5
- Every commercial page has one primary action named by outcome, a stated next step, and a lower-commitment alternative — and the page builds toward the ask.
Scores of 1 and 3 sit between the anchors above.
Evidence required
Name the primary action on your most important service page and quote its button text. If it reads “Learn More”, “Submit”, or “Contact Us”, score 3 or below.
CTA and Conversion Path: ___ / 5
Fix first: rewrite the primary button on your highest-traffic service page and add one line beneath it describing what happens next.
CHAPTER 05 · Tier 3 — Can they act?
Friction and Usability
Friction is what it costs someone to act after they've already decided to. This chapter is device-agnostic: broken specifically on phones belongs to Mobile, awkward everywhere belongs here.
Some friction is obvious — a broken form, a hidden phone number, a checkout that errors. Most of it is quieter: navigation organized around your internal structure, competing options where one would do, or a service page that never explains how to get started.
Each obstacle is small. They accumulate.
Follow the path yourself
Pick the most important conversion on your site and try to complete it. Not as a test — actually do it, start to finish, and count every:
- click
- field
- decision
- page load
- confusing instruction
- moment of uncertainty
Then ask of each whether it's necessary right now, or could wait until a conversation is underway. The exercise takes ten minutes and reliably surfaces what no amount of reading the page will.
Form friction
The first contact form is where most service businesses lose people, and almost always for the same reason: it's built around what the business wants to know rather than what's needed to start a conversation.
A first-contact form asking for
- Full name, company, job title
- Full mailing address
- Company size and annual revenue
- Budget range dropdown
- How did you hear about us?
- Detailed project history
- Preferred contact time and method
- A required 200-character description
What actually starts the conversation
- Name
- Email or phone — their choice
- One open field: what do you need help with?
- One line: we reply within one business day
You can gather everything else once someone is talking to you. A budget dropdown on first contact asks a stranger to disqualify themselves before they know what things cost.
Choice overload
More options don't reliably help. When a page presents six equally-weighted paths, the visitor has to decide which one matters before doing anything, and that decision is work you've handed them. The principle isn't that fewer buttons always convert better — it's that a page should know what it's for. Decide the page's job, let one action carry it, and demote the rest.
Diagnostic questions
- ✓How many clicks from homepage to completed primary conversion?
- ✓How many form fields does first contact require, and is each one needed now?
- ✓Is your phone number visible and clickable on every page?
- ✓Does navigation use customer language rather than internal structure?
- ✓Does any page present more than two meaningful options with equal weight?
- ✓Does every service page explain how to get started?
- ✓Are there dead ends — pages with no onward path?
Score yourself
0–5
Friction and Usability
- Score 0
- Something in the primary conversion path is broken — a form that errors, a dead link, a missing or non-functional contact method.
- Score 2
- The path works but is heavy: many required fields, several unnecessary steps, or navigation that requires guessing where things live.
- Score 4
- The path is short and the form asks only for what's needed. One or two unnecessary steps or competing options remain.
- Score 5
- Converting takes the minimum number of steps, the form asks only what's required to start a conversation, and every page offers an obvious onward path.
Scores of 1 and 3 sit between the anchors above.
Evidence required
State the number of clicks and the number of required form fields between your homepage and a completed inquiry. If you haven't walked it, score 3 or below.
Friction and Usability: ___ / 5
Fix first: delete every form field you don't need to start a conversation — the fastest change in this guide, and often the largest single reduction in friction available.
CHAPTER 06 · Tier 3 — Can they act?
Mobile Experience
Mobile isn't a smaller version of your desktop site. It's a different set of constraints, and a site can be technically responsive while being genuinely unpleasant to use on a phone.
Those constraints are physical rather than psychological. The screen is small, so less fits above the fold and long text blocks become walls. Input comes from a thumb with a far larger contact area than a cursor, so tightly-spaced targets get mis-tapped, and one-handed use puts the reachable zone in the lower half of the screen. Browsing happens in interrupted bursts and often in poor conditions — in a truck, in a parking lot, in bright sun. Overlays that are minor annoyances on desktop can cover most of a phone screen at once. It isn't that mobile visitors are less patient; the device imposes real costs the desktop version never charged.
Test on a real phone
Not a resized browser window. Group your check into three clusters:
Layout and viewport. Text size and line length, section spacing, headline wrapping, image cropping, and horizontal drift. Pay particular attention to that last one — if the page slides sideways at all, something is overflowing the viewport, usually a fixed-width element or an oversized image. It's a small bug that makes a site feel broken.
Tapping and navigation. Target sizes, spacing between adjacent tappable elements, whether the menu opens and closes cleanly, how many taps reach a service page, and whether your phone number actually dials when tapped.
Forms, keyboards, and overlays. Whether the right keyboard appears per field type, whether the field you're typing in stays visible when the keyboard opens, and whether cookie banners, chat widgets, and popups dismiss without precision tapping — or stack on each other and cover the call to action.
Diagnostic questions
- ✓Can someone understand the offer without zooming or scrolling?
- ✓Does the page ever drift horizontally?
- ✓Are primary buttons large enough to tap confidently with a thumb?
- ✓Does tapping the phone number start a call?
- ✓Does the correct keyboard appear for each form field?
- ✓Can every overlay be dismissed easily on the first attempt?
- ✓Is the primary action reachable in the lower half of the screen?
- ✓Are text blocks broken up enough to scan rather than read?
Score yourself
0–5
Mobile Experience
- Score 0
- The site is not usable on a phone — content overflows, text requires zooming, or the primary action cannot be completed.
- Score 2
- Technically responsive but awkward: cramped tap targets, horizontal drift, overlays covering content, or forms that fight the keyboard.
- Score 4
- Comfortable on a phone with a deliberate mobile layout. One cluster still has issues — usually overlays or form behavior.
- Score 5
- Designed for the device: readable without zooming, confidently tappable, forms that behave, and overlays that dismiss cleanly on the first try.
Scores of 1 and 3 sit between the anchors above.
Evidence required
Confirm you completed your primary conversion on a real phone, on cellular data, in a private window. If you have only checked a resized desktop browser, score 3 or below.
Mobile Experience: ___ / 5
Fix first: eliminate horizontal drift and enlarge the primary button's tap target. Both are usually small CSS changes with disproportionate effect on how finished the site feels.
CHAPTER 07 · Tier 3 — Can they act?
Speed and Technical Performance
Speed matters because of what it does to the sequence. A page that takes six seconds to show its headline has spent six seconds not communicating, and some visitors leave in that window having learned nothing about your business. Performance problems don't reduce your conversion rate directly — they remove people before your message gets a chance to work.
The three metrics worth knowing
Google's Core Web Vitals map onto three distinct experiences, each with a published “good” threshold assessed at the 75th percentile of real visits — meaning three out of four should meet it — and reported separately for mobile and desktop.
| Metric | What it measures | “Good” threshold |
|---|---|---|
| LCP (Largest Contentful Paint) | How long until the main content appears | 2.5 seconds or less |
| INP (Interaction to Next Paint) | How quickly the page responds when tapped | 200 milliseconds or less |
| CLS (Cumulative Layout Shift) | How much the layout jumps while loading | 0.1 or less |
Measure all three with Google PageSpeed Insights or Chrome's Lighthouse panel. Field data from real visits is more meaningful than a single lab test run from your office. Don't chase a perfect score for its own sake — a site at 90 that shows its headline immediately serves customers better than one at 100.
Symptom, cause, and fix
More useful than the metrics themselves is working backward from what you notice to what's causing it and who resolves it.
| What you notice | Likely cause | What resolves it |
|---|---|---|
| Headline or hero takes seconds to appear | Oversized hero image, or large background media such as an autoplaying video | Compress and correctly size the hero; serve modern formats; replace autoplay video with a poster image — developer |
| Text appears late, or flashes and re-renders | Render-blocking stylesheet, or fonts loading before text can display | Preload the primary font; match fallback font metrics; defer non-critical CSS — developer |
| Page jumps around while loading | Images without dimensions, injected banners, or fonts swapping in late | Set explicit width and height; reserve space for banners — developer |
| Buttons feel laggy after tapping | Too much JavaScript executing on load, usually from plugins | Remove unused plugins; defer non-critical scripts — developer |
| Gradually slower without a redesign, or fine on Wi-Fi and slow on cellular | Accumulated third-party scripts (chat, analytics, pixels, heatmaps) and total page weight | Audit and remove scripts that aren't earning their place; compress and size images per device — you can start this yourself |
That last row is the one to check first. Script creep and uncompressed images account for a large share of the slow sites we're asked to look at, and neither requires a developer to investigate.
Diagnostic questions
- ✓Have you measured LCP, INP, and CLS on real visits, not just a single lab test?
- ✓Does the main headline appear almost immediately on a phone using cellular data?
- ✓Does the layout stay still while the page finishes loading?
- ✓Do buttons respond immediately when tapped?
- ✓Do you know every third-party script running on the site, and why?
- ✓Is any image being served larger than it displays?
- ✓Does anything autoplay?
Score yourself
0–5
Speed and Technical Performance
- Score 0
- Pages take many seconds to show meaningful content on a phone, or fail to load reliably at all.
- Score 2
- Content eventually appears but noticeably late, the layout shifts while loading, or taps feel delayed. No measurement has been done.
- Score 4
- Measured, and mostly meeting the thresholds. One metric falls short, or performance is good on Wi-Fi but poor on cellular.
- Score 5
- All three Core Web Vitals meet their thresholds on real-visit field data, on mobile, and the main content appears almost immediately.
Scores of 1 and 3 sit between the anchors above.
Evidence required
State your LCP, INP, and CLS figures and where you measured them. If you have not measured, score 3 or below regardless of how fast the site feels to you.
Speed and Technical Performance: ___ / 5
Fix first: measure before changing anything, then compress the hero image and remove third-party scripts you can't justify — most of the easily recoverable time lives there.
CHAPTER 08 · Tier 4 — Can you tell?
Measurement
If you can't measure the actions that matter, you can't tell whether any of the previous seven chapters worked — you'll be making changes and forming impressions, and impressions are shaped heavily by the last conversation you happened to have.
Measurement sits last in the dependency chain but should be set up first in practice. Put tracking in place before you start repairing, or you'll finish the work with no way to evaluate it.
Define what actually counts
“Conversions” is used loosely enough to be misleading. These are genuinely different events with genuinely different value, and collapsing them into one number hides more than it shows:
- Form submission — someone sent a message. Includes spam and tire-kickers.
- Phone click — someone tapped your number. It doesn't confirm the call connected.
- Completed call — an actual conversation happened.
- Booked consultation — a scheduled commitment, usually a much stronger signal.
- Qualified lead — a real prospect, in your service area, for work you do, with a plausible budget.
- Sale — money.
Forty form submissions and two qualified leads is a different problem from four submissions and three qualified leads. The first has a targeting problem; the second works and simply needs volume. A single “conversions” number can't distinguish them, which is why defining your terms matters more than the tooling.
Pick your primary metric. For most service businesses it's qualified leads, which means somebody has to categorize inquiries as they arrive — a five-minute weekly habit, not a software purchase.
A minimum event set
For a typical service business, this is enough:
- ✓Contact form submissions, separated by which form
- ✓Phone number clicks (tel: link taps)
- ✓Booking or scheduling completions, if you use a scheduler
- ✓Primary CTA clicks on your main service pages
- ✓Which page the visitor was on when they converted
- ✓Traffic source for each conversion, at least at the channel level
- ✓A simple manual tally of which inquiries turned out to be qualified
That last one is the most valuable and the most often skipped. It's also the only one that requires no technical setup at all.
Why small samples mislead
The gate ranges apply here too, and the arithmetic is worth stating plainly. Three inquiries last month and five this month is a 67% increase — or it's two more phone calls, well within ordinary variation. Both descriptions are accurate; only one is useful.
That doesn't mean small businesses can't improve their websites. It means the evidence should match the claim. Fixing a broken form is justified by the form being broken. Concluding that a new headline caused a 40% lift, off a base of five, is not supported by five. Be patient with the numbers and decisive about obvious defects — those aren't in conflict.
Score yourself
0–5
Measurement
- Score 0
- No analytics, or analytics installed but never examined. Conversions are known only through impression.
- Score 2
- Sessions and pageviews are tracked, but no conversion events are. You can see traffic and cannot see outcomes.
- Score 4
- Key conversion events are tracked and attributable to pages and sources. Lead quality isn't distinguished from raw volume.
- Score 5
- Conversion events are tracked, attributed to page and source, and reviewed regularly — including a record of which inquiries were actually qualified.
Scores of 1 and 3 sit between the anchors above.
Evidence required
State how many qualified leads the site produced last month and which pages they came from. If you cannot answer from data, score 3 or below.
Measurement: ___ / 5
Fix first: track form submissions and phone-number taps, then start keeping a manual tally of which inquiries were qualified. The tally is worth more than any additional tooling you could buy.
Calculate Your Score
Add your eight scores. Traffic isn't among them — that was the gate.
01 Message Clarity ___ / 5
02 Offer Strength ___ / 5
03 Trust and Proof ___ / 5
04 CTA and Conversion Path ___ / 5
05 Friction and Usability ___ / 5
06 Mobile Experience ___ / 5
07 Speed and Technical Performance ___ / 5
08 Measurement ___ / 5
─────────
TOTAL ___ / 40
Before reading the bands, circle every category at 2 or below. Those override your total. The number is useful for orientation, but your two lowest categories are almost always more actionable.
0–12 · Foundational problems
Multiple areas are failing at once and compounding. Repairs at this level tend not to hold, because each fix depends on something else that's also broken. Don't increase ad spend yet. Start with Clarity and Offer, and be honest about whether the current site can carry the changes — if several categories scored 0 or 1, a rebuild is often cheaper than an extended sequence of patches.
13–24 · Leaking in several places
The site functions and loses people steadily. There's usually no single catastrophe; several moderate weaknesses are stacking, and any one alone wouldn't warrant much concern. Take your three lowest scores, order them by tier, and work through them one at a time.
25–32 · Working, with specific gaps
The fundamentals hold, and losses are coming from identifiable places rather than the site's basic design — which makes this the most efficient band to improve from. Watch for a blocker hiding inside a decent total: a 30 with Trust at 2 is a Trust problem, not a refinement opportunity.
33–40 · The website may not be your constraint
If you scored honestly, the site is likely no longer the main thing holding you back. That's a real finding, not an invitation to keep optimizing.
Look upstream instead: traffic quality and volume, whether the offer matches what the market wants, and the economics underneath it — what a customer is worth, what you can afford to spend acquiring one, whether your pricing supports the business you want. Further website work here tends to produce small returns while the actual constraint goes unexamined. One caution: a 33+ from generous self-scoring is worse than an honest 20, because it sends you looking in the wrong place.
What Should You Fix First?
Not the easiest problem, and not the one you find most annoying. The one furthest upstream.
- Find your blockers — every category at 2 or below.
- Fix them in dependency order, not worst-score-first: anything actually broken, then Clarity, Offer, Trust, CTA and Path, Friction, Mobile, Speed. A slightly worse Speed score does not outrank a Clarity blocker.
- With no blockers, take your two or three lowest categories and apply the same ordering.
- Change one meaningful thing at a time. Two simultaneous changes teach you nothing about either.
- Gather enough evidence before concluding it worked — the gate ranges apply, and conversion events matter more than sessions.
- Re-score. The diagnostic is meant to be run more than once.
Measurement is the exception to the ordering. It's last in dependency terms but should be set up before you start, since it's what lets you evaluate everything else.
The principle underneath all of it:
- ✓Score all eight categories using the evidence rule
- ✓Circle every category at 2 or below
- ✓Set up conversion tracking before changing anything
- ✓Fix one thing at a time, and note the date
- ✓Re-score once enough conversion events have accumulated
A Worked Example
Hypothetical composite example. The business below is a teaching illustration, not a WebCraft Labz client. The scores and figures are invented to demonstrate how the diagnostic resolves — they are not measured results, and no outcome is being claimed.
The business: a residential HVAC company, two crews, a metro area of about 400,000 people. The site is four years old, built on a template by a previous marketing vendor, and looks reasonably professional. Around 900 visits a month from local search, producing roughly eight inquiries — low, in the owner's view, for the traffic.
Gate: traffic is local and commercial-intent, mostly searches like ac repair near me, comfortably over 300 qualified visits. Outcome C — proceed.
Scores:
01 Message Clarity 2 / 5 ← blocker
02 Offer Strength 2 / 5 ← blocker
03 Trust and Proof 3 / 5
04 CTA and Conversion Path 3 / 5
05 Friction and Usability 2 / 5 ← blocker
06 Mobile Experience 4 / 5
07 Speed and Technical Performance 4 / 5
08 Measurement 1 / 5 ← blocker
─────────
TOTAL 21 / 40
Behind those scores. The homepage headline reads Your Comfort Is Our Priority — which doesn't say whether the company installs or repairs, serves homes or businesses, or answers emergency calls, the most decision-relevant fact in the trade. No pricing signal appears anywhere, the contact form asks for eleven fields, and analytics tracks no conversion events, so "eight inquiries" is a rough count from an inbox. Mobile and Speed score well: the template was competently built.
Why the total misleads. 21/40 lands in leaking in several places, which reads as "several moderate improvements needed." The blocker rule tells a different story — four blockers, two of them upstream in Tier 1. This isn't a site needing broad refinement; it's a site with a specific upstream failure and competent downstream work sitting on top of it.
Why it's still a repair, not a rebuild. Four blockers would trip the rebuild threshold on a raw count. But all four are content or configuration problems — a headline, a pricing signal, a form length, a tracking setup — and the platform isn't preventing any of them. Nothing here needs replacing; it needs writing and configuring. That's the distinction the tipping rule below is drawing.
Repair order:
- Measurement first, out of sequence — track submissions and phone taps, and start a manual tally of qualified inquiries. Nothing below can be evaluated without it.
- Clarity — rewrite the headline to name the service, the area, and emergency availability.
- Offer — add a pricing framework, and describe what happens after a call.
- Friction — cut the form to three fields and make the phone number tap-to-call sitewide.
- Trust and CTA — both scored 3. Address once the blockers are cleared and measurement can evaluate them.
- Mobile and Speed — leave alone. Nothing meaningful to recover.
Step 6 is the instructive one. Mobile and Speed are the site's highest scores, and a performance-optimization proposal would spend the budget there while the headline still fails to say what the company does. The most technically satisfying work available is also the work with the least to gain.
Repair, Redesign, or Rebuild?
Not every underperforming website needs replacing. Sometimes a few focused changes are enough, and knowing which situation you're in prevents both wasted money and wasted years.
Repair the existing site if
- the structure is sound and the platform is still maintainable
- mobile and performance problems are correctable rather than architectural
- the pages match the services you currently sell
- most of your blockers are in Clarity, Offer, Trust, or CTA — messaging and placement
Consider a redesign if
- the underlying system works but the site looks dated enough to undermine trust
- navigation needs restructuring around how customers actually think
- page layouts can't support your current offers, or conversion paths need real rework
- the business has evolved past what the site describes
Consider a rebuild if
- the platform actively prevents the fixes this diagnostic identified
- the site is expensive or fragile to maintain, or can't support integrations you now need
- mobile usability or performance problems are structural rather than cosmetic
- fixing the existing system would approach the cost of replacing it
The tipping rule
Read it off your blockers:
Blockers in Clarity, Offer, Trust, CTA, or Measurement are usually repairs. These are content, placement, and configuration problems. A competent site can carry better messaging without being rebuilt — and rebuilding won't fix them by itself, since you'd arrive at a new site with the same unclear headline.
Blockers in Friction, Mobile, or Speed may be either. The question is whether the platform permits the fix at reasonable cost. A form you can't shorten, a layout that won't behave on phones, or performance limited by the platform's architecture points toward rebuild. Fixable in place means repair.
Three or more blockers the platform won't let you fix, or a total below 13, usually points to rebuild — not because the number is magic, but because that many simultaneous structural failures normally means the foundation is the problem, and sequential patching tends to cost more than replacement within a couple of years.
Count the blockers the two rules above put in the rebuild column, not your raw blocker total. Four content blockers sitting on a sound platform is a long repair list, not a rebuild.
If you're weighing this, the project planner is built for exactly this decision — it maps scope before you commit to a direction.
Why We Aren't Giving You a Benchmark Conversion Rate
You've probably seen guides that answer this with an industry average — service businesses convert at X%, so if you're below that, you have a problem. We're not going to, and the reason is worth stating.
A conversion rate is a ratio, and both halves vary enormously between businesses that look similar on paper. What counts as a conversion differs, as the previous chapter covered. Traffic composition differs — a business with mostly branded search converts far higher than one with mostly informational traffic, with no difference in website quality. Price point, sales cycle, service area, and seasonality all move it. Two roofing companies in the same city can differ by several multiples for reasons unrelated to either website.
A benchmark without that context misleads in both directions: it tells some businesses they're fine when they're leaving money on the table, and others they're broken when they're performing normally for their traffic mix.
More useful is your own rate over time, measured consistently, with a clear definition of what you're counting. Your last quarter is a better benchmark than someone else's industry.
How This Guide Differs From Our Others
Three related guides, three genuinely different questions:
This diagnostic identifies what is broken and what to prioritize — score, find blockers, sequence repairs.
Marketing Websites That Actually Convert covers how a marketing website should be structured and built. Read it once you've identified what to fix and want to know what good looks like.
How Much Does a Custom Website Cost in 2026? answers the budget and scope question, for when the diagnostic points toward redesign or rebuild.
Diagnose here, design there, budget there. If you're starting from scratch rather than diagnosing an existing site, the structural guide is the better entry point.
Key Takeaways
- “My website doesn’t convert” describes at least eight distinct problems, and they don’t all deserve equal attention.
- Fix in dependency order. Improving downstream areas while an upstream blocker remains produces little change.
- Traffic is a gate, not a score. If the wrong people are arriving, or too few of them, the website diagnostic can't tell you much.
- Score above 3 only where you can point to specific evidence on a specific page.
- Any category at 2 or below is a blocker and outranks your total score.
- Your two lowest categories are more actionable than your total.
- Set up measurement before making changes, or you won't be able to tell what worked.
- A high score is a real finding: it means the constraint is probably traffic, offer, or economics — not the website.
Your Next Step
What to do depends on what the diagnostic found.
If your blockers are structural — three or more that the platform won't let you fix, or a total under 13 — you're likely in redesign or rebuild territory. Map the scope before committing: start your project plan.
If you have specific, targeted problems on an otherwise sound site, those are repairs worth doing in order. Website development and optimization services covers that work, or use the repair order above and do it yourself. Both are legitimate; the sequence matters more than who executes it.
If you scored 33 or above with no blockers, you probably don't need website work right now. Look at traffic quality and volume, offer-market fit, and the economics underneath both. We'd rather tell you that than sell you a rebuild you don't need.
Unsure which of the three you're in? Send us your scores and the URL, and we'll tell you straight — including if the answer is that the website isn't your problem.
Build Something Better
Ideas deserve a better home than a generic blog.
WebCraft Labz builds fast, polished websites for businesses and creators who want their ideas to feel premium from the first scroll.
Start a ProjectContinue Reading
