Your problem is not the amount of traffic.
It is that almost none of it could ever buy.
We joined your Search Console data to your Analytics enquiry data and looked at every page on the site. What came back changes what we think you should build, and it is not what anyone expected.
Three findings, in the order they matter
All of this comes from your own Search Console and Analytics, joined page by page across September 2025 to March 2026. We used that window deliberately, and the first finding explains why.
Any other agency looking at Wirehouse is working from third party crawl estimates, because that is all they can see from the outside. Those tools put your organic traffic at roughly 660 visits a month. Your own Search Console puts it between 1,300 and 2,100 clicks a month. The tools are out by about half, in your favour, and every proposal built on them starts from a number that is wrong.
We host your back end, we have your Analytics and Search Console, and we built the site that is there now. So this document is not built on estimates of your traffic. It is built on your own enquiry events, joined page by page to your own search data. That is also why the first finding is an admission rather than a sales point. When you have the real data you have to report what it says, including the parts that reflect on us.
The analytics stopped working in April
Search Console shows Google sending you between 1,300 and 1,700 clicks a month right through the summer. Analytics recorded 228 to 345 total sessions in those same months. Before April, the two broadly agreed.
So the website did not go quiet. The measurement did. Every performance number anyone has looked at since April is wrong, and wrong in the direction that makes the site look dead when it is not.
We should have caught this sooner. We host the back end and we had the access, and it took us until this week to join the two data sets and notice. We would rather say that plainly than present six months of broken tracking as a discovery.
There is a second fault sitting underneath it. page_view is configured as a key event, so Analytics reports around 125,000 "conversions" against 43,000 sessions. Conversion rate is the metric you told us matters most, and it has been meaningless since last October.
Half your search traffic lands on pages that have never produced an enquiry
72 pages pull 6,492 search clicks and have generated zero enquiries. That is 53% of everything Google sends you, arriving on content that has never once turned into a conversation.
Meanwhile the homepage alone produces 71% of all enquiries. Only 30 pages produce an enquiry at all, and just 24 of those also earn search clicks.
Where your search traffic actually goes
Sep 2025 to Mar 2026, 12,356 clicks joined to enquiry events
The worst performers are all blog posts, and they are doing well in search: workplace housekeeping at 604 clicks, nut allergies at 414, managing performance issues at 363, witness statements at 291. Genuinely useful writing, ranking properly, read by people who will never buy anything.
The channel you have invested least in is already your best
Organic search is 34% of your sessions but 68% of your enquiries. It converts at 2.12% against direct at 0.35%. We rebuilt these figures from the underlying enquiry events rather than the broken key event, which is why they still hold. That is six times better, from the channel with the least money behind it.
This is the good news, and it is the whole commercial case. The thing that works is the thing nobody has been feeding.
One number we will not overclaim
Search Console anonymises low volume queries, so only 3,587 of those 12,356 clicks have a query attached, which skews the visible sample towards brand terms. We have quoted the page level join throughout, because that covers every click. Where we cite query percentages, treat them as directional.
Benchmarked against the market you actually compete in
Organic performance for the providers you named, plus WorkNest shown for reference, pulled from Ahrefs for Great Britain in September 2026. This is the gap the retainer is designed to close.
| Provider | Domain rating | Organic keywords | Monthly organic traffic | Free employer tools |
|---|---|---|---|---|
| BrightHR | 70 | 2,209 | 202,915 | 13 |
| Citation | 64 | 1,967 | 26,024 | 5 |
| Croner | 60 | 3,915 | 21,431 | 6 |
| Peninsula | 64 | 3,019 | 18,754 | 2 |
| Avensure | 54 | 999 | 15,331 | 11 |
| WorkNest | 67 | 575 | 5,286 | 0 |
| Wirehouse | 36 | 165 | 662 | 0 |
Ahrefs, Great Britain, September 2026. Tool counts verified by opening each provider's tools directly. WorkNest is shown for reference as part of your wider group, not as a competitor.
The number that matters most in that table
Citation's holiday entitlement calculator is the second most visited page on their entire website, at roughly 2,900 visits a month. That single page pulls more than four times the organic traffic of all of wirehouse-es.com.
Croner's maternity pay calculator is their top content page, at nearly double their own homepage. Avensure's working hours calculator outperforms their homepage too. Free tools are not a marginal tactic in this market. They are how the category is being won, and Wirehouse currently has none.
Worth being straight about one thing: several of those particular tools are searched by employees as much as employers. They prove that tools win the category, not that any tool will win you enquiries. We are deliberately recommending the employer-only subset, which is lower volume and much higher intent.
The gap, in specific terms
Keywords where at least one of these providers ranks in the top ten and Wirehouse does not rank at all. Filtered to difficulty 25 or under, because at domain rating 36 that is what is genuinely reachable.
| Search term | Monthly searches | Difficulty | Who holds it today |
|---|---|---|---|
| risk assessment template | 12,000 | 1 | Citation #1, Peninsula #6 |
| bradford factor calculator | 3,300 | 2 | a DR 33 site, with a DR 1 site at #7 |
| fire risk assessment template | 3,500 | 3 | no commercial provider in the top five |
| hr advice line | 350 | 3 | Citation, Peninsula and Avensure, all top three |
| hr services for small business | 450 | 2 | Croner #1, Avensure #3 |
| health and safety consultants for small businesses | 350 | 0 | Avensure #3, Citation #8 |
| chas accreditation | 1,200 | 0 | a DR 14 and a DR 35 site |
| health and safety consultants construction | 100 | 17 | a DR 6 site at #2 |
The whole gap set runs to roughly 860 keywords. These are the ones where the difficulty is low enough to win from where you are now.
Why this is winnable rather than wishful
Avensure sit at domain rating 54 and hold top-three positions on several of the commercial terms above. That matters because it proves these results are not bought with domain authority alone. The pages winning chas accreditation include a domain rating 14 site. The page at number seven for bradford factor calculator has a domain rating of 1. At 36, Wirehouse is not the underdog in these particular fights.
Everything on the site was written for the wrong person
This is the single idea that connects all three findings, and it is the reason a redesign alone would not fix anything.
Look at what actually ranks. Workplace housekeeping. The difference between a hazard and a risk. Nut allergy risk assessments. Section 7 of the Health and Safety at Work Act. Some of these sit at position one and pull hundreds of clicks a month.
They are read by students revising for NEBOSH and IOSH, employees checking their rights, and people who are curious. Not one of them runs a business with 25 to 75 staff and a problem they need solving this week.
We made the same mistake ourselves, last week
The first set of tools we prototyped for you included a redundancy calculator and a holiday entitlement calculator. They looked excellent on paper. Redundancy calculator alone is 35,000 searches a month.
Then we checked who searches it. Overwhelmingly employees, working out their own payout. We would have built you three more pages that pull traffic and produce nothing, which is exactly the problem you already have 72 times over. So we took them out and rebuilt them.
We are telling you this because it is the clearest possible illustration of the trap, and because we would rather you saw us catch it than discover it in six months.
The filter we would apply to everything from now on
One question, asked before a single page or tool gets built.
Would only an employer open this?
If an employee might search it too, the traffic will look good and the enquiries will not come. Bradford Factor, yes. Redundancy calculator, no.
Does it map to something you sell?
Every asset should sit directly in front of one of your three revenue lines, not float in a general resources section.
Is the next step obvious?
A page that answers a question and then stops is a page that produces readers. The answer and the route to you have to arrive together.
Three things to click
Working prototypes in your own brand. Type into them.
The page a buyer should land on, with the gaps annotated
Built to current best practice and wired to your own words from the 23rd. Leave the annotation toggle on and every block explains what is there now and why it is changing. This is the one to open first.
Open the service page →Absence & Bradford Factor, Tribunal exposure, HSE Fee for Intervention
No employee has ever calculated their employer's Fee for Intervention liability. Each tool maps to one service line, and the statutory rate table is exposed so your team updates it each April without a developer.
Open the tools →A homepage that asks for the enquiry
Your existing brand in a cleaner layout, with a working tool in the hero so the first thing a visitor can do is get an answer rather than read three paragraphs about who you are.
Open the concept →There are two versions of your website, and the machines get the worse one
This is the finding that connects the framework question to the thing you actually care about. We checked it this week by fetching your pages the way a crawler does, then again in a real browser, and comparing.
| Page | Words in the HTML your server sends | Words after JavaScript runs | H1 in the served HTML |
|---|---|---|---|
| /outsourced-hr-services/ | 303 | 1,646 | None |
| /health-safety/ | 586 | 745 | None |
| /employment-law/ | 792 | 885 | None |
| Homepage | 272 | 707 | Yes |
Fetched 25 September 2026. Word counts are the main content area only. The same 162,576 byte file is served to Googlebot, to a GPTBot user agent and to a desktop browser, byte for byte, so there is no dynamic rendering and no prerender fallback in place.
Why this is the AI visibility problem, not a technical footnote
Google runs JavaScript, so Google eventually sees the full page. GPTBot, ClaudeBot, PerplexityBot and CCBot do not. They read what the server sends and nothing more.
So when an AI assistant is asked about outsourced HR support, the version of your HR page it can read is 303 words with no heading, no packages and no FAQ. Not the 1,646 word page you paid for. You told us an AI referral is as valuable as a normal referral, and that this is why you wanted to invest in the area. This is the first reason you are not getting more of them, and it is upstream of every piece of content anyone could write for you.
Look at the last column again. The homepage is the only page on the site with a heading in the served HTML, and the homepage produces 71% of your enquiries. We are not claiming that is the whole explanation. We are saying it is the one structural difference between the page that works and the pages that do not, and it has never been tested.
This is our build. We chose Gatsby, we shipped it, and the pages have been serving without their headings ever since. We are not presenting it as a discovery about someone else’s work.
Four more things we found while we were in there
All verified on the live site this week. Two of them are ours to fix and we would fix them in the first fortnight whatever you decide about the rest of this document.
The Fee for Intervention figure is wrong
/health-safety/ quotes HSE Fee for Intervention at £174 per hour. The rate has been £188 since 1 April 2026, and £183 before that from April 2025. The page has understated a client’s exposure by 8% for around eighteen months, on the page whose entire product is staying current with regulation. Fix this one this week.
A phone number that does not dial
The hero call button on /outsourced-hr-services/ points at tel:0333215005. That is ten digits. The other two service pages carry the correct eleven. Anyone tapping it on a mobile gets an invalid number. Same page, first sentence, reads “legally complaint” rather than compliant, twice.
No structured data anywhere on the site
Zero JSON-LD blocks on every page we checked, before and after rendering. No Organization, no LocalBusiness, no BreadcrumbList. No canonical tags and no Open Graph tags either, so anything shared into Teams, LinkedIn or WhatsApp renders without a title or image. /about-us/, which your navigation links to, returns a 404.
One Trustpilot review
Your profile has been claimed since October 2021 and holds 3.7 from a single review, left in May 2024. Peninsula has 23,684 at 4.9, Citation 8,228, Croner 7,082, Avensure 1,971. Seer Interactive studied 804,491 AI responses this March: brands with no Trustpilot presence were cited about 1% of the time, brands with even a handful of reviews 53.5%. Reviews are now an AI visibility asset, not just a trust one.
You cannot yet tell a referral from an AI referral from a buyer
This one came out of your own answers rather than the data, and it is the reason the investment is hard to justify internally today.
“We only get around 5-10 enquiries per month. Some that come through the website are already aware of us or have been recommended us, some have come via AI, only 2-3 probably come because they’ve found us and are intent buyers.”
Read that again with a commercial eye. Five to ten enquiries arrive, and the split between word of mouth, AI and genuine search discovery is an estimate. Asked which resources have produced business, the honest answer was “don’t know”. That is not a criticism, it is what happens when tracking has been broken since April and there has never been a source field on the form.
It matters more than usual here, because you also told us this:
“Referrals are always the highest converting. What we have seen is an ‘AI referral’ is as valuable if not more valuable than normal referrals. This is why we wanted to explore investing further in this area.”
You want to invest in AI visibility because AI referrals convert. You currently have no way to count one. That is the gap we would close first, because without it you will be asked to renew this budget in twelve months on a feeling rather than a number.
Where your enquiries actually sit today
Rebuilt from the underlying enquiry events across the clean months, September 2025 to March 2026. Note the middle row.
| Channel | Sessions | Enquiries | Rate | What it really contains |
|---|---|---|---|---|
| Organic search | 15,019 | 318 | 2.12% | Your best understood channel, and still 68% of enquiries from 34% of sessions |
| Direct | 23,709 | 82 | 0.35% | The blind bucket. Word of mouth, returning clients, typed URLs, and an unknown share of AI assistant traffic that arrives with no referrer |
| Referral | 1,004 | 33 | 3.29% | Your highest converting channel by rate, and the smallest. Nobody has ever looked at what is in it |
| Paid | 476 | 11 | 2.31% | Small volume, not a factor either way |
Your GA4, enquiry events only (form_submit, phone_click, mail_click), excluding the broken key event and excluding May 2026 as a data anomaly.
What we would put in place, in the first fortnight
None of this is expensive. All of it is the difference between renewing this budget on evidence and renewing it on hope.
- An AI assistant channel group of its own. ChatGPT, Perplexity, Copilot, Gemini and Claude split out of Direct and Referral so you can finally see the AI referrals you already believe in, and watch them grow as the visibility work lands.
- A source question on the form. The only way to catch a word of mouth referral, because they arrive as Direct and always will. One field, asked well, and suddenly the biggest bucket has a shape.
- Aggregator traffic tagged separately. You told us you use aggregators alongside the website. Today they are mixed in. Separating them is what lets you answer the question you actually care about, which is how much less reliant on them you are getting.
- Source carried into the enquiry itself. Channel, landing page and first search term attached to every submission, so the sales team see where a lead came from at the point they pick up the phone, and so cost per enquiry can be reported by source rather than in aggregate.
- The key event configuration fixed, so conversion rate, your number one KPI, means something again.
Do this first and every number in the rest of this document becomes checkable by you rather than assertable by us. That is deliberate. We would rather be held to a real measurement than protected by a broken one.
Everything that follows answers something you said
You answered ten questions on 23 September and said you were keeping it short. We have taken the answers literally rather than loosely, so here is where each one lands in this document.
| What you said | Where it is answered |
|---|---|
| “Clearly defined service proposition pages that allow enquiries to be made easily. We want this moving in weeks not months.” | The sitemap leads on your three retained lines, and the first service pages ship on your current site in weeks, before the rebuild starts |
| “The main service pages will be HR and Employment Law Retained, Health and Safety Retained and Health and Safety Consultancy.” | That is the exact top level of the proposed sitemap. We have not invented a different architecture |
| “Tribunal Support packages which includes an indemnity called the Advice Guarantee.” | Pulled out as the lead differentiator on the service page mockup. It is currently a line of body copy |
| “We want social proof. We’ve not been great at using our clients… helping with that would be significant.” | We run the case study programme, positioned as ongoing problems solved rather than one off incidents, which was your framing |
| “A closed deal for HR is likely to be more profitable, however the HR market is more saturated.” | Why the strategy wins the long tail rather than fighting head on, and why the H&S consultancy work is positioned as the low commitment way in |
| “An AI referral is as valuable if not more valuable than normal referrals.” | The attribution work above, so you can count them, and the AI visibility work in the retainer, so there are more of them |
| “Most are contacted within 10 mins of making an enquiry.” | Noted and left alone. Response time is not your problem and we have not padded the proposal by pretending it is |
| Which resources have produced business: “Don’t know.” | Finding 02. We joined the two data sets and answered it. The answer is almost none of them |
| “An initial investment followed by ongoing monthly investment for 12 months… that ongoing investment may cover some of the front ended build.” | Offered exactly that way in the commercials, as the second of the two structures |
| “If we don’t double the number of opportunities, the website hasn’t justified the investment.” | The commercial section is tested against that bar, and reports the result honestly even where it falls short |
What this is, in one line
This is not a website project. It is a lead generation engine for a business that already has demand it cannot see and traffic it cannot convert. Right now there are gaps at every stage: the measurement is broken, the pages that rank cannot sell, the pages that could sell do not exist, and the enquiries that do arrive cannot be traced back to what produced them.
We close those gaps with the four things in order. Data first, so decisions stop being opinions. Then design and UX, so the pages a buyer lands on are built to produce an enquiry rather than to describe a service. Then SEO and GEO, which compounds, because every page, tool and case study keeps earning long after it ships and every one of them is an asset you own rather than a lead you rent.
The sitemap
916 URLs live today. 325 earn a search click and only 30 produce an enquiry. This is what we would keep, rebuild, add and retire. Click any section to open it.
Comparison content, if you want it
There is real demand for honest side by side comparison pages in this market, and they are exactly the kind of content AI engines quote when someone asks which provider to use. We have left them out of the structure above because they need a decision from you rather than from us.
If you do want them, we would build them on verifiable, attributable facts such as published contract terms and what is included as standard, and we would avoid superiority claims entirely. That keeps the content useful and keeps it defensible.
The framework, and whether to go headless
You asked us directly what to do about the platform. Here is the honest answer, including the part that makes this cheaper than you are probably expecting.
First, a correction worth making early. You are already headless. The site runs WordPress on WP Engine as the content back end, with a Gatsby front end on Netlify reading from it. So the question is not whether to move to headless. It is whether to replace the front end, or both halves.
That matters commercially, because replacing just the front end is a substantially smaller project than the full replatform people usually imagine when they hear the word rebuild.
We are not going to tell you Gatsby is dead
It is not, and anyone who says so can be contradicted in about a minute. The project is still maintained, downloads are steady, and Netlify said in 2023, when Gatsby Cloud closed, that Gatsby itself was not being sunset.
The problem is Gatsby 4 specifically, and it is documented by the Gatsby project itself rather than by us.
Version 4 left support in Q4 2023
That is from Gatsby's own published version support table, not our interpretation of it.
The project marks 4.x as unsupported
Their own SECURITY.md lists 5.x as supported and 4.x as not. In practice that means no security patches.
You are on 4.25.6, published March 2023
Roughly three and a half years old, and no release of any kind on the 4.x line since May 2024.
Your exact version sits on a published advisory
It affects the local development server rather than your live site, so this is a hygiene signal rather than a breach. We would rather be precise than alarming.
So there are two honest routes, and we will show you both
Upgrade in framework, Gatsby 4 to 5
Cheaper and faster. Puts you back into a supported version and stops the security question. It does not change what the site does, and it leaves you on a framework where feature development has effectively stopped.
Replace the front end
More work now, and it removes the question for the next several years. It is also the only route that lets us build the conversion layer, the tools and the structured data properly rather than bolting them onto an ageing build.
Our recommendation, and the reason for it
Replace the front end with Next.js, and keep the WordPress back end your team already knows. Not because Gatsby is bad, but because everything in this proposal that produces enquiries needs to be built rather than restyled, and doing that on a version that stopped receiving security patches three years ago would be building on ground we know is moving.
If budget is the constraint, the Gatsby 5 upgrade is a legitimate smaller step and we will say so rather than pretend it is not an option.
The services the group now delivers
HR Consultancy, payroll, HR admin, HR training and GDPR now sit elsewhere in the group. The pages are still on your site and they still attract search demand. You told us the decision on what happens to those leads is still being debated internally.
We are not going to wait for that decision, and we are not going to guess at it either. We would build the routing so that it works either way, and so switching between the two is a setting rather than a rebuild.
If the group takes the lead
The page keeps the traffic, captures the enquiry, and hands over cleanly with the context attached. It becomes a referral line rather than a dead end, and you can see how much it is worth.
If you re-cost internally
Same page, same capture, the enquiry comes to Wirehouse and the delivery is arranged behind the scenes. The visitor experience is identical.
Either way
The content stays, the rankings stay, and nothing is thrown away while the decision is made. Today those visitors reach a page for something you no longer sell and nothing happens at all.
What we need from you first
You gave us the list on the call and it was genuinely useful. Before we design the routing we want to confirm it is complete, because if GDPR moved then it is worth checking nothing else did quietly. That is a twenty minute conversation, not a workstream.
Our process, and what you get at each stage
Click through it. Nothing gets built until the thing before it is signed off, which is how we avoid the redesign that looks lovely and changes nothing.
Discovery
We already did most of this before you asked us to. The analytics join, the page by page audit and the search demand work in this document are all discovery output. What remains is the part only you can give us: what you actually sell, what a good lead looks like, and where the group is going.
- Definitive service list, what you sell versus what the group has taken
- Fix the analytics and the key event configuration, so we measure the same thing
- Lead definition and what happens after a form is submitted
- Content approval workflow, because that is your real bottleneck
- Brand horizon conversation, openly
At this stage
Sitemap & content model
We agree the structure before anyone draws anything. Every page gets a job, a search intent and a next step, or it does not make the map. This is also where we decide what happens to the roughly 590 URLs that are currently diluting the site, and where the services the group now delivers get routed.
- Full sitemap, signed off page by page
- Keep, rebuild, consolidate or retire decision on every existing URL
- Redirect map drafted before a single page moves
- Routing design for services you no longer deliver
- Content model so your team can publish without a developer
At this stage
Wireframes
Structure and hierarchy in grey boxes, deliberately without colour or imagery, so the conversation is about what goes where and in what order rather than whether you like the blue. This is the cheapest possible moment to change your mind, and the stage most agencies skip.
- Wireframes for every page template, desktop and mobile
- Where proof sits, where the enquiry route sits, on every template
- Tool placement and how a tool hands over to a conversation
- Two rounds of revision built in
At this stage
Design
Now it gets visual. Your brand, brought current, built as a system rather than a set of pictures, so the twentieth page looks as considered as the first. You have already seen a homepage direction, so we are not starting from a blank page.
- Design direction on key screens, three concepts
- Full design system: type, colour, components, states
- Every template designed at desktop and mobile
- Three rounds of revision
- Accessibility checked as we go, not audited at the end
At this stage
Build & test
The current site stays live and untouched throughout. You get a staging environment from week one of build, so you are watching it come together rather than waiting for a reveal. Structured data, performance and crawlability are built in at this stage, not retrofitted, because during a rebuild they cost almost nothing and afterwards they cost a lot.
- Front end rebuilt, CMS wired, content migrated
- Schema, llms.txt and technical SEO at build time
- Forms consolidated onto an API service with instant confirmation
- Every redirect tested individually before launch
- Structured UAT round with your team
- Launch with rollback available
At this stage
Iterate
Launch is the start of the useful part. Now the tracking works, we can see which pages produce enquiries and which do not, and build accordingly. This is where the retainer lives, and it is the difference between a website that was good on launch day and one that is better every quarter.
- Monthly reporting on enquiries and cost per enquiry, not rankings
- New tools added in priority order, highest intent first
- Case studies produced on a rolling basis, we run the process
- Content against the commercial gap, not the traffic gap
- AI visibility tracked and optimised
At this stage
Prove it on the current site, then rebuild on the evidence
You said weeks not months. This is how both things are true: the fixes that produce enquiries do not need the rebuild, so they ship first.
Fix the measurement
Repair tracking, remove page_view as a key event, un-orphan the 19 whitepapers and 29 webinars that no page links to. Cheap, fast, and we both start seeing the truth.
Ship the conversion layer
Three service pages and the first two employer tools, live on the current site. First new enquiries arrive here, before the rebuild has started.
Rebuild
Sitemap, wireframes, design, build. Informed by what the first two phases have already taught us about what converts.
Compound
Retainer. Tools, content and case studies in priority order, reported on enquiries and cost per enquiry.
Why this order protects you
You told us the brand may not exist in one or two years, depending on what the group decides. We have taken that seriously in the sequencing rather than ignoring it.
Everything in the first five weeks produces enquiries on your current site and moves with you regardless: content, tools, rankings, captured audience and case studies are all portable. The rebuild, which is the part most tied to this brand, comes after you have seen the earlier work produce results. If the answer from the group changes, you will have spent the smallest amount at the point of greatest uncertainty.
This costs more than the plan you already have, and here is exactly why
You have a twelve month plan from us dated 3 September at £23,000. You said on the 7th that it was still under review, and then the call on the 23rd changed the brief. Rather than send you a bigger number and hope you do not compare them, here is the comparison.
What changed between the two
The September plan moved you off Gatsby like for like. Same design, same content, same structure, and it said in writing that we would not produce content. That was the right answer to the brief at the time. It is not the answer to the brief you gave us on the call.
Held only if this is signed by 30 September. Weeks 1 to 14. Invoiced 50% at kick off, 50% at launch.
- Tracking repaired and the key event configuration corrected, inside the first fortnight
- Three commercial service pages and the first two employer tools shipped on your current site, before the rebuild starts
- Full replatform of the front end off Gatsby 4 onto a supported framework
- Existing content migrated, URLs preserved, redirects tested
- Homepage and key templates redesigned around enquiry rather than description
- Structured data, llms.txt and technical SEO built in rather than retrofitted
- Form consolidation and instant confirmation
- CMS training, documentation, and an approval workflow your legal and brand sign off can live with
Includes the hosting and maintenance you pay £250 a month for today, so the true increase is £2,050. Rolling monthly, no lock in, 30 days notice.
- The commercial keyword layer, built out in winnable order
- Further employer tools, highest volume and lowest difficulty first
- AI visibility work, so the engines that already quote you start quoting you to buyers rather than to students
- The case study programme run by us, since collecting them internally has not happened
- Content produced and shepherded through your approval workflow
- Reporting against commercial enquiries first, cost per enquiry second, organic traffic third
The rebuild as scoped is an £18,000 piece of work and that is what it will be quoted at from October. We would rather land it in this quarter than next, so the number is £14,999 if you can get it signed off by 30 September. That is the whole reason, and it is ours rather than yours. Nothing in the scope changes either way, and if the group needs longer than five days, tell us and we will talk about it rather than pretend the deadline is a cliff.
On the monthly, you told us in July that £3,000 was too much for the ongoing, and that the heavy lifting sits up front and maintenance should be lighter. You were right, and the number has come down to £2,300 even though the scope has gone up. The intensity is deliberately concentrated in the first five weeks and carried by the one off fee, which is where it belongs.
Checked against your numbers, not ours
Every input below is yours, from the KPI form and the call. Correct any of them and the conclusion moves with you.
You wrote £10,000 in your email and £9,000 on the KPI form. We have used the lower of the two.£9,000
Nothing lands in months one and two. The first new enquiries arrive around week five.about 3.5x
You set the bar at five to ten times TCV. On a conservative ramp, year one does not clear it. We would rather show you that than flatter the model, because you will run this arithmetic yourself and we would rather you trusted the next number we give you. It passes five times from around month ten, and sits near the top of your range in year two, for the simple reason that the build is bought once and the asset keeps earning. If a twelve month window is the only one that counts internally, tell us now and we will cut scope to fit it rather than quietly miss it.
Two ways to structure the same total
As quoted
£14,999 across the build, then £2,300 a month for twelve months. Cleanest if the build sits in a capital line and the retainer in marketing.
Your structure, from the call
£8,999 at kick off, then £2,800 a month for twelve months. The same £42,599, with more of the build amortised into the monthly, which is what you asked for.
Commercial terms
The retainer is rolling monthly, no lock in, 30 days notice, and it absorbs your current hosting and maintenance agreement rather than sitting on top of it. We recommend twelve months to let the compounding show, and we would rather you left at month five than signed a term you resent. Invoicing in week three, payable at the start of the following month. Delivered work is non refundable. The one off is invoiced 50% at kick off and 50% at launch. All figures exclude VAT.
The two decisions are genuinely separable. You can take the one off on its own and decide on the retainer once the first five weeks have produced enquiries. We would not recommend the reverse, because the retainer compounds against a conversion layer that does not exist yet. And if the group review lands somewhere that makes the rebuild the wrong call, the September plan at £23,000 is still a perfectly good answer to the smaller question. We would rather sell you that than sell you this badly.
Why us, given we built the current one
A fair question, and worth answering directly rather than hoping it does not come up.
We know this codebase, we host the back end, and we already have the analytics access. That means no discovery tax, no three weeks of an agency working out how your site is put together, and no handover risk. It also means that where something on the current site is our doing, we have said so in this document rather than presenting it as a discovery.
“Saigon Digital transformed how Ski.com shows up online. Beyond rebuilding our platform, they helped us rethink our entire search and AI visibility strategy. We saw a significant uplift in organic traffic, our content started appearing in AI-generated recommendations, and the quality of inbound leads improved dramatically. They understand where digital is heading and how to turn it into real commercial results.”
Years combined experience
Across development, search and AI visibility.
Continue beyond the first project
Most of our work is with clients we already have.
Countries served
With a delivery team in one place, not scattered across freelancers.
The measurement first. Then the conversion layer. Then the rebuild.
Everything in this document came out of your own data. The next step is a short session to agree what you actually sell, so we can design the routing properly.
The commercials are set out in full above, including how they compare with the plan you already have. The one thing we still need from you is the decision on the group service list, because it changes the routing rather than the price.
Book the next session → Nick Rowe · CEO & Co-Founder, Saigon Digital · nick@saigon.digital