Blog

Analytics platforms present hundreds of metrics. Almost all of them are irrelevant to a business with one location and a phone.

Here is the short list that matters, the list you can ignore, and one question that outperforms all of it.

The question that beats any dashboard

Ask every new customer how they heard about you. Write down the answer.

That is it. That is the highest value measurement available to a small business, it costs nothing, and it captures things no analytics tool can see.

Analytics cannot tell you that somebody was recommended by a neighbour, saw your van, or remembered a job you did on their street four years ago. For a lot of local businesses those channels dominate, and they are entirely invisible in any dashboard.

Do it for three months, consistently, on every enquiry. Then count. You will almost certainly find that your customers come from somewhere other than where you assumed, and that finding usually changes where the effort should go.

The failure mode is inconsistency. If you ask half the time, the data is useless. Put it on the form, add it to the intake script, make it a field in whatever you use to track jobs.

The numbers worth watching

Enquiries per month. Calls, forms, messages, walk ins. One number, tracked monthly. This is the closest thing to a headline metric a small business has, and most do not track it at all.

Where they came from. From the question above.

Conversion from enquiry to customer. Of the people who got in touch, how many bought. If this is low, your marketing is working and something after the enquiry is not, which is a completely different problem and one that spending more on marketing will make worse.

Customer value. What a customer is worth over the time they typically stay with you, not just the first transaction. You are the only person who can supply this and it is what makes every other number meaningful.

Website visitors, roughly, and how many are on phones. You do not need precision. You need to know whether it is fifty a month or five thousand, because that determines whether website changes can plausibly matter.

Your Core Web Vitals. Specifically Largest Contentful Paint on mobile, against Google’s published 2.5 second threshold. Check it quarterly in PageSpeed Insights. It is free and it takes a minute.

Review count and recency on Google. How many, and how recent. Recency matters more than most people assume.

That is the whole list. Seven items, most of them checkable in a few minutes a month.

What to ignore

Bounce rate. Frequently misunderstood. A visitor who arrives, reads your phone number and calls you registers as a bounce. That is a success recorded as a failure.

Time on page. Ambiguous in the same way. Long could mean engaged or confused.

Impressions. How many people theoretically could have seen something. Interesting to nobody with a business to run.

Follower count. Vanity, unless your business model is genuinely audience based. A thousand followers who will never buy from you are worth less than eleven local customers.

Likes. Discussed elsewhere on this site. Not customers.

Keyword rankings in isolation. Ranking first for something nobody searches is not an achievement. Enquiries are the measure.

How to measure a change honestly

When you fix something and want to know whether it worked, three rules.

Change one thing at a time, or accept that you will not know which change did it. This is frustrating and it is how measurement works.

Give it long enough. For most small businesses, two weeks of data is noise. A month is minimum, a quarter is better, and if your enquiry volume is small, longer still.

Compare like with like. Same period last year, not last month. A landscaper comparing March to June has learned about spring, not about their website. This single error invalidates more small business conclusions than anything else.

The simplest system that works

You do not need software. A spreadsheet with one row per month and these columns:

Enquiries. Customers. Where they came from, as a tally. Anything notable that changed.

Fill it in on the first of the month. It takes ten minutes. After a year you have something genuinely valuable that no analytics platform could have given you, because it includes the offline sources that dominate local business.

If you want website numbers as well, install something simple and check it monthly. Do not check it daily. Daily numbers on a small business site are almost entirely noise and watching them produces anxiety rather than insight.

The honest limitation

Some things cannot be measured cleanly and you should not pretend otherwise.

Brand perception. Word of mouth. The cumulative effect of looking like a real business for years. Whether the person who chose your competitor did so because of your website or because their brother in law was available.

Marketing measurement for a small business is directional, not precise. Anybody offering exact attribution at your scale is overselling, and decisions made on false precision are worse than decisions made on honest estimates.

What you can know is whether enquiries went up, where people say they came from, and whether the things you fixed are still fixed. That is enough to steer by.

Start here

If you do nothing else from this article: ask every new customer how they heard about you, and write it down, for three months without exceptions.

That single habit will teach you more about your marketing than any dashboard, it costs nothing, and it works for exactly the channels that matter most to a local business and that no software can see.

Two free tools worth connecting

You do not need paid analytics, but two free things are worth the twenty minutes it takes to set them up.

Google Search Console. Shows what people actually searched before your site appeared, how often they clicked, and which pages get seen. This is real query data from real people, which is better than any keyword research, and a startling number of businesses have never connected it.

Your Google Business Profile insights. Shows how people found you, whether they searched your business name or your category, and what they did next. Name searches mean your reputation is doing the work. Category searches mean you are being discovered by strangers. The ratio between them tells you something useful about which part of your marketing is functioning.

Both are free, both take about ten minutes to connect, and both answer questions no amount of guessing will.

The review you should do quarterly

Once every three months, sit down for half an hour with your spreadsheet and ask four questions.

Did enquiries go up or down compared to the same quarter last year? Same quarter, not last quarter.

Where did customers say they came from? Has the mix changed?

What did I change in the last three months, and did anything move?

What is the one thing I would fix next?

That is the whole review. It is enough to steer a small business and it does not require any expertise, any software, or any consultant.

Why precision is the wrong goal

A closing thought, because the pressure to measure everything is strong and it is often unhelpful at this scale.

Large companies invest heavily in attribution because a one percent improvement on enormous volume is worth real money. At your scale, the effort required to measure precisely usually exceeds the value of the precision.

What you need is enough information to make the next decision. Are enquiries trending up or down. Where do people say they find me. Did the thing I fixed appear to help.

That is a low bar and it is genuinely sufficient. Businesses that chase perfect measurement tend to spend more time in dashboards than with customers, which is a poor trade at any size and a terrible one at yours.

Ask people how they heard about you. Write it down. Everything else is refinement.

Setting a baseline before you change anything

If you are about to fix your website, start asking for reviews, or begin posting regularly, spend twenty minutes first recording where you are today.

Enquiries in the last three months. Current review count and the date of the most recent one. Your mobile PageSpeed score and LCP figure. Roughly how many people visit your site monthly.

Write those five things down with today’s date.

This costs almost nothing and it is the difference between knowing whether something worked and guessing. The most common frustration I hear is somebody who invested in their presence and cannot tell whether it helped, and it always traces back to no baseline existing beforehand.

You cannot go back and measure the past. Twenty minutes now buys you the ability to answer the question in six months.

The one metric to watch if you only watch one

Enquiries per month, compared to the same month last year.

That single line, tracked consistently, will tell you more about whether your marketing is working than any dashboard, and it accounts for the seasonality that makes every other comparison misleading.

Everything else on this page is refinement on top of that one number.

---

Core Web Vitals thresholds, including the 2.5 second Largest Contentful Paint target measured at the 75th percentile, are published by Google at web.dev.

The Baymard Institute maintains a running meta analysis of cart abandonment research, collecting every credible study rather than running one and publishing a single figure. As of their September 2025 update, the average across 50 documented studies sits at 70.22 percent.

The number worth paying attention to is not really the average. It is the spread. The individual studies range from roughly 55 percent to 84 percent, and that range is the honest part, because it tells you abandonment varies enormously with what is sold, to whom, and how.

So roughly seven in ten people who add something to a cart do not buy it, and that has stayed broadly stable for years despite everything the industry has thrown at it. Some of that is unavoidable. People browse, compare, save things for later. But a meaningful share is caused by things on the page, and those are fixable.

Here is what a product page has to do.

1. Photographs that answer questions

Product photography is not decoration on an e-commerce page. It is the entire substitute for handling the object, and it is doing the job that a shop assistant and a shelf do in person.

Multiple angles, including the unflattering ones. People want to see the back, the underside, the seams. Hiding them creates suspicion, not mystery.

Scale. The single most common photography failure. A person cannot tell how big something is from an isolated shot on white. Show it in a hand, next to a familiar object, or being used. Returns caused by size surprises are expensive and almost entirely preventable here.

Detail shots. Texture, finish, the quality of construction. This is where you demonstrate that it is worth the price.

In context. The item in use, in a real setting. Isolated product shots on white are necessary and insufficient.

Consistent treatment. The same lighting and background across your catalogue. A store where every product was shot differently looks improvised.

2. A description that answers the actual question

Most product descriptions are written as marketing copy. They should be written as answers.

What is it made of. What size, precisely, in numbers. How does it behave over time. What is included. What is not included. How do I look after it.

Read your customer service messages and note what people ask before buying. Every one of those questions belongs on the page. If somebody had to ask, the page failed, and for every person who asked there were several who left instead.

Specificity outperforms enthusiasm consistently. “Eleven inches across, holds about a litre” does more than “generously sized.”

3. Price clarity, including the total

Baymard’s own checkout research points repeatedly to unexpected costs as a leading cause of abandonment, and it is worth taking seriously because it is one of the few causes entirely within your control.

The failure pattern is familiar: a customer builds a mental total through the process, then shipping and fees appear at the final step and the number jumps. That feels like a bait and switch even when nothing dishonest happened, and people leave over the feeling rather than the amount.

Show shipping cost, or the threshold for free shipping, on the product page. Show taxes if they apply. Show delivery timeframe. Every figure the customer will eventually see should be visible before they commit attention to checking out.

4. Stock and delivery, stated plainly

In stock or not. How long until it ships. How long until it arrives. Where you ship to.

Vagueness here is expensive because it is one of the last uncertainties before purchase. “Ships in 2 to 3 business days” is a small sentence that resolves a real hesitation.

5. Returns, said clearly and early

A clear returns policy is a conversion tool, not a legal footnote. It removes the main risk of buying something unseen.

Say it plainly and near the buy button. “30 day returns, we pay postage” does more work than any amount of persuasive copy, because it transfers risk away from the customer at the exact moment they are weighing it.

Burying it in a footer link works against you, because the anxiety is present whether or not the answer is easy to find.

6. Reviews on the product itself

Site wide testimonials do very little on a product page. Reviews of this specific item do a great deal.

If you sell clothing or anything with fit, reviews that mention sizing are worth more than everything else on the page combined. Prompt for that specifically when you ask.

A handful of genuine reviews including a lukewarm one is more persuasive than a wall of perfect scores, which reads as curated because it usually is.

7. A checkout that does not fight the customer

Baymard’s usability research consistently identifies forced account creation as a major abandonment cause. Somebody who is ready to give you money should not be required to invent a password first.

Offer guest checkout. Keep the form short and ask only for what you need to fulfil the order. Show progress if it runs to multiple steps. Do not clear the form when validation fails, which is a small failure that produces disproportionate rage.

8. It has to load

Everything above is irrelevant if the page does not appear. Product pages are especially vulnerable because they are image heavy, and image heavy pages are where oversized files do the most damage.

Google’s threshold is 2.5 seconds for main content, measured at the 75th percentile. Run your best selling product page through PageSpeed Insights on mobile and see where you stand. If images are named as the cause, they can almost always be a fraction of their size with no visible difference.

The diagnostic

Take your best selling product page and go through this list, marking each item present or absent.

Then buy something from your own store on a phone, all the way through to payment, as if you were a stranger. Notice every point where you hesitate or have to hunt for something. Those are the same points where customers leave, and you will find two or three you did not know about.

Then do the same on a competitor’s site and note what they answer that you do not.

Most stores are missing three or four items from this list. None of them require a rebuild, and the ones about clarity, shipping costs and returns are usually just writing.

The mobile version is the real version

Most of the traffic is on phones and most product pages are designed on a desktop, which produces predictable failures.

The buy button is below the fold and stays there. On a long product page, somebody scrolling through images and description reaches a point of decision and then has to hunt for the button. A sticky add to cart bar that stays visible while scrolling solves this and is standard on well built stores.

The images cannot be examined. On a phone, tapping an image should open it large and allow zooming. If it does not, you have removed the one thing that substitutes for handling the product.

The variant selectors are fiddly. Small dropdowns for size and colour, close together, are hard to use with a thumb. Larger tappable options are better and reduce mis-selection, which reduces returns.

The description is a wall. On a narrow screen, four paragraphs of unbroken text will not be read. Break it into short sections with headings so somebody can find the fact they want.

Buy something from your own store on a phone, all the way to payment. You will find at least two of these.

What to fix first

If you are going through this list with limited time, the order that produces the most improvement per hour:

Add shipping cost and delivery timing to the product page. Unexpected costs at checkout are a leading documented cause of abandonment and this is entirely within your control.

Enable guest checkout if you do not already. Forced account creation is one of the most consistently identified abandonment causes in Baymard’s usability research.

Add a scale reference to your main product photos. One image showing the item in a hand or beside something familiar. This reduces both hesitation and returns.

State the returns policy near the buy button. One sentence.

Check the page speed on mobile. Product pages are image heavy and this is where oversized files do the most damage.

Those five are mostly writing and settings rather than development. None of them require a redesign, and together they address the causes that research consistently identifies rather than the ones that feel most pressing.

What abandonment is not telling you

A closing caution about the 70 percent figure, because it gets used badly.

A high abandonment rate is not automatically a problem to be solved. Carts get used as wishlists. People compare across three stores with items sitting in each. Somebody adds something to check the shipping cost with no intention of buying today.

None of that is a broken checkout. It is normal browsing behaviour, and no amount of optimisation will convert most of it.

What you are actually looking for is the portion of abandonment caused by something on your page: a surprise cost, a forced account, a missing answer, a page that would not load. Those are recoverable and they are the ones worth chasing.

The way to find them is not the abandonment percentage. It is going through your own checkout on a phone and noticing where you hesitate, and reading the questions customers ask you before they buy. Both of those point at specific fixable things. The headline number points at nothing in particular, which is why it is quoted so often and acted on so rarely.

---

Cart abandonment figures from the Baymard Institute’s meta analysis of 50 documented studies, updated September 2025: average 70.22 percent, individual studies ranging from roughly 55 to 84 percent. Core Web Vitals thresholds published by Google at web.dev. This article does not estimate what abandonment costs any specific store.

Your site is not performing. Somebody suggests a rebuild. It is a large number and a long project, and you have no way to judge whether it is necessary.

Here is how to work it out yourself, before anyone quotes you.

Start with the diagnosis, not the treatment

The mistake is deciding on a rebuild before establishing what is wrong. A rebuild is a treatment, and prescribing treatment before diagnosis is how people spend a lot of money and end up with a prettier version of the same problem.

So first, get specific about the symptom. “It’s not working” is not a diagnosis. These are:

People arrive and leave immediately. People cannot find what they need. People find what they need but do not make contact. Nobody arrives at all. It looks unprofessional next to competitors. It cannot be updated without help. It is broken on phones.

Each of those points somewhere different, and only some of them are about the website at all. If nobody arrives, your problem is upstream of the site and a rebuild will not solve it.

The case for fixing

Repairs are the right answer more often than the industry admits, because the industry is paid for rebuilds.

Fix if the structure is sound and the problems are specific. Slow loading, an untappable phone number, a confusing homepage sequence, missing evidence, outdated content. These are afternoon jobs on an existing site.

Fix if it works on phones. Mobile capability is the dividing line. A site that is genuinely responsive can usually be improved. A site built before mobile mattered, that only works on desktop, generally cannot be patched into shape.

Fix if you can edit it. If you or someone in your business can update text and images without paying a developer, that is a functioning asset worth improving.

Fix if you are under two years old. Early businesses change what they sell. Building an expensive site around a proposition that will shift in eight months is a common and avoidable waste.

The case for rebuilding

Rebuild if it does not work on phones. Not “looks a bit cramped.” Genuinely broken. Sideways scrolling, unreadable text, buttons that cannot be tapped. Retrofitting responsiveness onto a site that never had it usually costs more than starting again.

Rebuild if nobody can edit it. A site where every text change requires a developer will not be maintained, and an unmaintained site decays into a liability regardless of how good it looked at launch.

Rebuild if the platform is dead or unsafe. Abandoned software, unpatched, unsupported. This is a security issue and not really a marketing decision.

Rebuild if the business genuinely changed. Different services, different market, different name. When the underlying proposition changed, the structure usually has to.

Rebuild if you have tried fixing and measured no change. Note the word measured. Fixed it, gave it a fair period, compared like with like, and nothing moved.

The question that resolves most cases

Ask this: if every specific problem I can name were fixed on the current site, would I be satisfied?

If yes, do not rebuild. Fix the specific problems. You have just described a repair job and the answer is not a new site.

If no, and you cannot articulate what else is wrong, be careful. That usually means the dissatisfaction is aesthetic or emotional rather than functional. Sometimes that is legitimate, because looking dated has a real cost when you are quoting against competitors. But it is worth being honest that this is the reason, rather than dressing it up as performance, because it changes what you should be buying.

Getting an assessment you can trust

The person telling you what to do usually stands to be paid for the answer. That does not make them dishonest, but it is a real bias.

Ask for the diagnosis separately. “Before we discuss a rebuild, tell me what specifically is wrong with the current site.” A good answer is a list of concrete issues. A vague answer about it being outdated is a warning sign.

Ask what a repair would look like and cost. If they refuse to consider repairs at all, that is information about them rather than about your site.

Get the free evidence yourself first. Run PageSpeed Insights. Do the phone test at evening brightness with one thumb. Have somebody who does not know your business try to find your prices while you watch. That gives you an independent picture before anyone quotes.

Ask what happens to your search presence. A rebuild done carelessly can lose accumulated ranking, because URLs change and redirects get missed. Anybody competent will raise this before you do.

The middle option nobody offers

There is a third path that rarely gets proposed because it is less profitable: a targeted rework.

Keep the platform, keep the structure, replace the homepage and the two or three pages that matter, fix the speed, replace the stock photography with real work, and leave the rest alone.

This is frequently the correct answer for a small business. It addresses the pages that actually get seen, costs a fraction of a rebuild, and takes weeks instead of months. If your site is fundamentally sound but the important pages are weak, ask specifically for this and see what comes back.

An order of operations

If you are unsure, do it in this sequence and stop as soon as things improve.

First, fix what is free. Speed, tappable phone number, correct details, real photos, homepage sequencing. Days of work, little or no money.

Then measure. Give it a fair period and compare against the same period last year, not last month, especially if your business is seasonal.

Then rework the key pages if the fundamentals are fine but the important pages are not doing their job.

Then rebuild, if you have done the above and the constraint is genuinely structural.

Most businesses that follow this order never reach step four, and the ones that do reach it arrive with a clear brief and evidence about what was actually wrong, which produces a much better rebuild than starting from a vague dissatisfaction.

The honest summary

Rebuilds are sometimes necessary and frequently oversold. The tell is whether anyone diagnosed the problem before proposing the treatment.

Do the free diagnosis yourself. Write down the specific failures. Then ask whether fixing that list would satisfy you.

Most of the time it would, and most of the time nobody asked.

What a rebuild actually involves

If you do decide on one, knowing the shape of the project helps you judge quotes and spot omissions.

Discovery and structure. What pages exist, what each is for, what the site is trying to achieve. Skipping this is how you end up with a prettier version of the same problems.

Content. Somebody has to write the words and supply the photographs. This is the step that delays most rebuilds, because businesses assume the designer handles it and designers assume the business does. Agree explicitly who is doing it before you start.

Design and build. The visible part, and usually the shortest.

Migration and redirects. Every old URL needs to point to its new equivalent. Get this wrong and you lose accumulated search presence, which is a genuinely painful and entirely avoidable loss. Ask about it specifically.

Testing. On real phones, on cellular, by somebody who is not the person who built it.

Handover. Can you edit it? Do you own the domain, the hosting and the accounts? Get answers in writing.

That last point matters more than it sounds. A rebuild that leaves you unable to change your own text, or unable to access your own domain, has replaced one dependency with another.

The question to ask yourself in a year

Whichever route you take, the outcome worth aiming for is a site you can and will maintain.

The most common cause of a bad website is not a bad build. It is three years of nobody touching it. Services changed, prices changed, staff left, photos aged, and nothing got updated because updating required a phone call and an invoice.

So when comparing options, weight this heavily: after this project, can somebody in my business change the text and swap a photo without paying anybody?

If the answer is no, you are buying something that will decay on a schedule, and you should factor that into the comparison alongside the price.

The seasonal timing question

One practical consideration that gets ignored and occasionally costs a lot.

Do not rebuild during your busy season.

A rebuild needs your attention: content, decisions, feedback, testing. If it lands during the eight weeks when your phone does not stop, one of two things happens. Either the project stalls because you cannot respond, or it ships without your proper review because you were too busy to look.

There is also risk in changing your site right before the period when it matters most. Migrations occasionally go wrong, redirects occasionally get missed, and the worst time to discover that is the week your customers are actively searching.

Rebuild in your quiet season, launch before the busy one starts, and give yourself a few weeks of buffer to catch problems.

---

This article contains no statistics and no pricing claims. Where numbers appear elsewhere on this site, they are measured with free public tools or cited to a named source.

I look at a lot of small business websites. The same five problems account for most of the damage, they are largely invisible to the owner, and none of them require a rebuild.

Here they are in order of how much they cost.

1. It is slow, and you have never noticed

Why you have not noticed: You look at your site on a laptop, on your own wifi, with every file already cached from the last time. On that machine it is instant and has been for years. Your customer is on a phone, on cellular, from a cold start, and having a completely different experience.

What it costs: Every visitor who leaves before the page appears was, by definition, interested enough to click. They are the most qualified people you have and they are lost before you get to say anything. Whatever you spent to earn that click still cost you.

How to check: Google PageSpeed Insights, mobile score only. Look at the Largest Contentful Paint figure. Google’s published threshold for good is 2.5 seconds, measured at the 75th percentile of loads.

The usual cause: Oversized images. Somebody uploaded photos straight from a camera and the browser scales them down to fit a phone. The picture looks identical. The load time does not.

Time to fix: Resizing images is often an afternoon. Deeper problems are usually days, not months.

2. The phone number is not tappable

Why you have not noticed: You have never tried to call your own business from your own website.

What it costs: This one is stark. A person on a phone, ready to call, taps your number and nothing happens. They now have to select the text, copy it, switch apps, paste it. Some will. Many will go back to the search results where the competitor’s number is one tap.

You lost somebody at the exact moment of highest intent, which is the most expensive possible place to lose them.

How to check: Open your site on a phone and tap the number. If the call screen does not appear, it is broken.

The usual cause: The number is plain text, or part of an image, or the link was written incorrectly.

Time to fix: Minutes.

3. Nobody can tell what you do or where you are

Why you have not noticed: You know what you do. When you look at your homepage, your brain supplies the context automatically. A stranger has none.

What it costs: A visitor who cannot immediately place you leaves. They are not going to work at it.

How to check: Open your homepage on a phone and look only at what is visible before scrolling. Can you answer three questions from that alone: what does this business do, where is it, how do I contact it?

The usual cause: The information exists, further down, below a large hero image and several paragraphs about commitment to quality. It is a sequencing problem, not a content problem.

Time to fix: An afternoon, and usually no new writing.

4. The evidence is stock photography

Why you have not noticed: It looked professional when the site was built, and stock images are designed to look acceptable.

What it costs: Most people recognise stock photography, at least subliminally. A generic image of a smiling person in a hard hat signals one of two things: you had nothing of your own to show, or you did not think it mattered. Both undermine you, and both are usually untrue, because you almost certainly have done good work worth photographing.

You also lose the actual benefit. A photo of your real work is evidence. A stock photo is decoration and carries no information.

How to check: Look at every image on your site and ask whether it shows your work, your premises or your people. Anything that does not is costing you.

The usual cause: No archive of real photos, because nobody was capturing as they worked.

Time to fix: Immediate for the swap. The underlying fix is a habit: one before shot and one after shot on every job, as a rule rather than a decision.

5. Your details disagree with each other

Why you have not noticed: You would have to look at three things side by side, and there is no reason to.

What it costs: Nobody investigates an inconsistency. They register a small flicker of doubt and move on slightly less confident, and slightly less confident loses when two other options are on the same screen.

There is a second cost. Search engines use consistency of business name, address and phone across sources as a signal of legitimacy, so mismatches work against you in local search as well.

How to check: Open your website, your Google Business Profile and your main social bio at the same time. Compare phone, address, hours and business name character by character. “Street” and “St” count as different.

The usual cause: Things got updated in one place and not the others. Old listings you forgot exist.

Time to fix: An hour, once you find them all. Search your own business name to turn up the listings you had forgotten.

The pattern underneath

Four of these five share a cause: you cannot see your own website the way a stranger does.

You are the worst possible tester of your own site because you know what everything means, where everything is, and what the business does. Every ambiguity that would stop a visitor is invisible to you, because your brain fills it in before you notice a gap existed.

Which is why the most useful thing on this entire list is not any individual fix. It is this: hand your phone to somebody who does not know your business, ask them to find out what you do and get in touch, and then watch without helping.

You will learn more in ninety seconds than from any amount of analysis. It is also mildly excruciating, which is why almost nobody does it.

What this adds up to

None of these five is a redesign. Together they are perhaps two days of work, most of it unskilled, and they address the failures that decide whether an interested person becomes an enquiry.

Business owners tend to assume that if their site is underperforming, they need a new one. Usually they need five specific things fixed on the one they have, and the reason they do not know that is that nothing about this problem announces itself.

Go and check the five. Write down which ones apply. That list is the actual size of the problem, and it is almost always smaller than the version in your head.

Three more worth checking

Not in the top five, but common enough to mention and all quick to fix.

Contact forms that go nowhere. Forms break silently. The address they send to gets closed, the plugin updates, messages start landing in spam. Send yourself a test message through your own form today, and then put a reminder in to do it every few months. Businesses have lost months of enquiries to this and only discovered it by accident.

A footer full of links to dead social accounts. You are advertising the abandoned page problem. Remove links to anything you no longer maintain.

No indication of service area. If you travel to customers, say where. A visitor who cannot tell whether you cover their town assumes you do not.

Why these five specifically

The five are not the most serious problems a website can have. They are the ones with the worst ratio of damage to effort.

Every one is invisible to the owner, cheap to fix, and positioned at a point where an interested person is deciding. That combination is what makes them expensive. A genuinely bad website is a known problem that somebody eventually addresses. These sit undetected for years in businesses that believe their website is fine.

The pattern to take away is not the specific list. It is that the failures costing you most are the ones you are structurally unable to see, which is why the outside test matters more than any checklist including this one.

Hand your phone to somebody who does not know your business. Watch without helping. Ninety seconds.

The order to fix them in

If all five apply, do them in this sequence, which is by effort rather than by impact.

Today, in under an hour: make the phone number tappable, correct any details that disagree across your website, Google profile and social bios.

This week, one afternoon: rearrange the homepage so what you do, where you are, and how to contact you are visible before scrolling. Replace the worst stock photos with real ones if you have any.

This month: resize the images and clear out unused scripts and widgets. If somebody else maintains your site, this is a specific request rather than a vague complaint about speed, which makes it far more likely to get done.

Ongoing: start capturing photographs of your work so the stock photography problem never returns.

The first two groups cost nothing but attention and they address the failures at the point of highest customer intent. Most businesses could clear them in a single Saturday and never do, because nothing about these problems announces itself.

---

Google’s Core Web Vitals thresholds, including the 2.5 second Largest Contentful Paint target measured at the 75th percentile, are published at web.dev. This article deliberately does not estimate what any of these failures costs a specific business, because that number depends on your traffic and your customer value, which only you have.