Citation-friendly pages
MISSINGZero of 80 fetched URLs match best-of, alternatives, comparison, review or pricing patterns. Nothing on the site is shaped the way answer engines quote.
Your site barely shows up on Google for the searches your audience is running. Here's what's broken, and the 15k+ monthly searches you could rank for instead.
Diginow is a UK web design and digital agency with studios in London and Leicester, building bespoke CMS, recruitment and eCommerce sites, plus AI-powered SEO, hosting and managed Microsoft 365.
The funnel is a free website audit and a discovery call.
The work is real: Elemis, Headspace, CFP Energy and LJB Recruitment all have case study pages, and the services page carries schema and 2,044 words. Search doesn't know any of that.
Our index finds one ranked keyword for the domain in the top 100, "pave recruit" at position 6, and the URL ranking for it sits on test.diginow.co.uk rather than the live site. Ranked keywords are down 92.3% over the last six months.
Two things need fixing before content matters: kill the staging subdomain from the index, and start building link signals, because an agency selling SEO that ranks for one keyword has a trust problem with buyers as well as with Google.
8 structural issues.
Diginow runs a well-built agency site with almost no search presence. 94 sitemap URLs, 80 pages fetched all returning 200, every entry carrying a lastmod, newest content 7 days old. Zero noindex pages, zero missing canonicals, zero cross-domain canonicals. Content depth is real: case studies run 1,200 to 1,450 words on a consistent Challenge, Our approach, The Solution, Key Results structure, and the homepage and services page both clear 2,000 words. Schema is present on most templates.
Against that, our index found 1 ranked keyword for the whole domain: "pave recruit" at position 6, 140 searches a month, about 4.73 in estimated traffic value. A 94-page agency site with 12 case studies and 53 editorial posts earning one ranked term is a targeting and indexation problem, not a writing problem. The single term it does rank for is a client's brand name, and the ranking URL sits on a staging subdomain. Commercial service pages (9 URLs under /service/) and the entire editorial library (44 under /insight/ plus 9 under /insights/) return zero ranked keywords between them. Genuine strengths worth protecting: clean technical base, recognisable client names (Elemis, Headspace), and case studies with actual results sections. Those assets are currently pointed at nothing a buyer searches for.
The one keyword the domain ranks for, "pave recruit" at position 6, resolves to a URL on test.diginow.co.uk, not the live site. A staging host serving 200s to Google means every page on the production site has a full duplicate elsewhere. Google picks one version per duplicate cluster, and when the staging copy wins, the live page inherits none of the equity. It also means any link earned to that case study is pointing at a host you will eventually turn off, at which point the ranking disappears entirely.
1. Confirm how many test.diginow.co.uk URLs are indexed with a site: query and in Search Console.
2. Put HTTP basic auth on the entire staging host so Googlebot cannot fetch any page there.
3. 301 every test.diginow.co.uk URL to its live www.diginow.co.uk equivalent before locking it, so existing equity transfers rather than dying.
4. Submit a removal request for the staging host in Search Console to speed up de-indexation.
5. Add a deploy-time check that fails the build if a non-production host is missing auth.
The 112-character homepage title, "Web Design London & Leicester | Bespoke Website Design Agency London | Recruitment, eCommerce & AI SEO | Diginow", is served verbatim on /get-a-free-website-audit and /book-a-discovery-call. Worse, the 170-character homepage meta description appears on every one of the 12 case studies sampled and on /contact. The deterministic crawl counts 2 duplicated titles and 2 duplicated metas across the fetched set, and the case-study pattern shows the template is falling back to a site default rather than generating per-page text. Google reads title and description as the page's own topical claim.
1. Fix the CMS SEO template so title and meta description default to the page H1 plus a per-template suffix, never to the homepage values.
2. Write unique titles under 60 characters for /get-a-free-website-audit and /book-a-discovery-call, leading with the action ("Free Website Audit for UK Businesses | Diginow").
3. Write a unique 140-155 character meta description for each of the 12 case studies, naming the client, the sector, and one result figure.
4. Re-crawl after deploy and confirm the duplicate title and duplicate meta counts both drop to zero.
The /service/ prefix holds 9 sitemap pages and returns 0 ranked keywords. The /services hub page (2,044 words, schema present) also returns 0. These are the money pages: bespoke CMS, recruitment website design, eCommerce website design, AI-powered SEO, hosting and support. The business sells discovery calls off them and Google is not showing them for anything. The mechanism is that the pages are written as capability descriptions rather than answers to a search. The /services hub uses H2s like "Applied AI for digital growth" and "Hosting, support & maintenance", which are internal service names, not queries.
1. Assign one primary keyword to each of the 9 /service/ pages from the pillar list, starting with "recruitment website design" and "ecommerce website design".
2. Rewrite each page H1 and title to match that keyword in plain buyer language, dropping internal service naming.
3. Add a pricing or cost band section to each service page, since "website design cost uk" is a top pillar and no service page currently addresses cost.
4. Link each service page from the two or three case studies in its sector using descriptive anchors.
5. Track ranked-keyword count per /service/ URL monthly and escalate any page still at zero after 90 days.
The sitemap carries 44 URLs under /insight/ and a further 9 under /insights/, plus an /insights hub. Both prefixes return 0 ranked keywords. Two things are wrong. First, the split means internal links, crawl attention and any earned authority are divided between two parallel trees that Google has no reason to treat as one section. Second, at 53 posts and one ranked keyword sitewide, the editorial program is producing pages that match no query. The hub page itself is 2,992 words with 172 outbound links and an H1 of "Pinboard & findings", a label no one searches and which tells the crawler nothing about the topic of the section.
1. Pick one canonical prefix (/insights/) and 301 every /insight/ URL to its equivalent, updating internal links rather than relying on redirects.
2. Pull the 53 posts against impressions data and archive or 410 anything with zero impressions in 6 months.
3. Rewrite the hub H1 from "Pinboard & findings" to a keyword-bearing heading such as "Web design and SEO insights for UK businesses".
4. Rebuild the remaining posts into two clusters, cost and process, each linking up to the matching /service/ page.
5. Cap new publishing until each new post is assigned a target keyword with real search volume before it is written.
On /our-work, /insights, /book-a-discovery-call and all 12 sampled case studies, the extracted H2 list includes "This site uses cookies to store information about your computer.", "Strictly Necessary" and "Analytics". The consent banner is being written into the document body as real heading elements rather than as an overlay outside the content flow. Google parses the heading hierarchy to understand what a page is about, so on /our-work half the visible H2 structure is consent boilerplate rather than portfolio content.
1. Move the consent banner markup outside the main content container and render its labels as div or span elements, not h2.
2. Confirm the change by re-fetching /our-work and /insights and checking that no consent text appears in the heading outline.
3. While in that template, change the /our-work H1 from "Portfolios" to a descriptive heading such as "Web design case studies for UK brands".
4. Add breadcrumb and ItemList schema to /our-work and /insights, which currently return no schema at all.
Case-study titles average well over 120 characters. The Eurowines page carries a 185-character title, Mander Centre 151, RJC Group 138, RFID Technologies 134. Each is the client's positioning sentence pasted into both the H1 and the title tag. Google truncates SERP titles around 60 characters, so what a searcher actually sees is a fragment of a sentence with no indication of what Diginow did. There is no sector term, no service term, and no "case study" label in the visible portion.
1. Rewrite all 12 case-study titles to the pattern "Client name recruitment website design case study | Diginow", capped at 60 characters.
2. Keep the long client sentence as an on-page subheading below the H1, and shorten the H1 to the client plus the project type.
3. Add the sector and service words (recruitment, eCommerce, bespoke CMS) to every case-study title so the pages can support the matching /service/ page.
4. Add CreativeWork or Article schema with the client name to each case study.
5. Re-check SERP title rewriting rates after 30 days.
The homepage H1 is the literal string "New Home page", a CMS placeholder left in from a build. The H1 is the single strongest on-page statement of what a page is about, and the site's most authoritative URL is using it to say nothing. The title tag and meta on that page are targeting web design in London and Leicester, recruitment sites, eCommerce and AI SEO, so the H1 contradicts every other signal on the page instead of reinforcing it. The homepage carries 2,178 words, 130 links and valid schema, meaning the rest of the page is doing its job while the headline element is blank.
1. Replace the homepage H1 with the primary commercial claim, for example "Bespoke web design for recruitment and eCommerce brands".
2. Confirm there is exactly one H1 on the page after the change.
3. Sweep every template for other placeholder strings ("New", "Untitled", "Lorem") before publishing.
4. Add an editorial check to the CMS publish flow that blocks a page going live with a default H1.
Image alt coverage across the 80 fetched pages is 54 percent. On a portfolio-led agency site this matters more than average: the case studies and /our-work are largely visual, and the screenshots of client sites are the proof of capability. An image with no alt attribute is invisible to Google Images and contributes nothing to the page's topical signal, so the strongest evidence Diginow has (real builds for Elemis, Headspace, Saville Row) is being shown to humans and hidden from the crawler. Image search is also a realistic secondary channel for a design agency, where buyers browse visually before they read.
1. Audit which pages fall below the 54 percent average and prioritise /our-work and the 12 case studies.
2. Write descriptive alt text naming the client and the project type, for example "Elemis eCommerce homepage designed by Diginow".
3. Make the alt field required in the CMS media uploader so new images cannot ship empty.
4. Re-measure coverage after the sweep and hold the site above 95 percent.
3 pages analysed for AI-writing patterns.
Scored 65/100 across 3 pages. The content reads AI-heavy: formulaic structure, multiple categories of tells (19 flagged), and no consistent editorial voice. A core update is likely to demote these pages.
“visibility is being shaped not only by ranking performance, but by”
Textbook "not only X but Y" scaffolding used to manufacture depth.
“B2B search is no longer just about where your page ranks.”
"not just X" opener setting up the pivot clause. Second instance of the same construction on one page.
“It means using headings and structure well. It means reducing duplication, improving page quality”
Three consecutive "It means..." sentences in one short paragraph. LLM anaphora a human editor would collapse.
“This is not the moment to chase gimmicks. It is the moment to sharpen fundamentals”
Negation-then-affirmation mirror sentence. Sounds decisive, carries no information.
“become clearer, more trusted and more citeable in answer-led search”
Rule-of-three abstract-noun list, repeated across the page ("expertise, relevance and trust", "authority, trust and commercial growth").
“The goal is not to game the system. It is to become the best available answer.”
Aphoristic closer with zero specifics. Standard LLM article sign-off.
“failing to create enough relevance, confidence or direction once the right people arrive”
Triadic abstract-noun list with no concrete referent. First of several rule-of-three constructions stacked through the piece.
“A strong B2B website aligns audience, message and journey more effectively.”
Second rule-of-three in the same shape. LLM cadence: three abstractions joined by 'and', zero specifics.
“the pages, journeys and positioning that turn visibility into opportunity”
Third identical triad. A human editor would vary the rhythm; the model repeats the template.
“traffic looks healthy, but qualified leads remain inconsistent”
'Traffic' and 'qualified leads' recur in nearly every paragraph as full noun phrases where a pronoun or synonym would be natural.
“In other words, traffic does not automatically become opportunity.”
Restates the previous paragraph adding no information. Classic LLM paragraph-closing echo.
“When these issues combine, the result is predictable. Visitors arrive, but too few move forward in a meaningful way.”
Short declarative-then-restate pairing with over-hedged filler ('in a meaningful way'). Voiceless assertion with no data, example or named case.
“That is a more useful and more commercial question. It shifts the focus from volume to effectiveness.”
Stock consultant-LLM pivot: pose 'the better question', then declare the abstract shift. No evidence follows.
“This is not just about simplification for its own sake. It is about protecting high-intent demand”
Textbook "not just X, but Y" scaffolding used to manufacture depth where no new information is added.
“the process becomes too awkward, too slow or too unclear”
Rule-of-three adjective stacking with no specifics. LLM default rhythm.
“Each point of friction reduces momentum. The more moments of hesitation a journey creates, the more likely it is that candidates will leave.”
"friction" appears four times across short adjacent paragraphs where an editor would vary or cut.
“It captures what matters without overloading the user. It works well on mobile. It explains what happens next.”
Four consecutive subject-verb sentences all starting "It" — machine cadence, not written prose.
“it is tempting to assume the role was not a good fit. Sometimes that is true. Often, it is not.”
Stock LLM concession-and-pivot opener that sets up a false objection to knock down.
“A better recruitment website does more than attract candidates. It helps them act.”
Empty aphoristic closer with zero evidence or specificity. Standard LLM paragraph capstone.
100% of sampled rankings come from searches that don't contain the brand name. That's traffic SEO earned, not reputation.
Already ranking close to page 1. These move first: a sharper title, a content refresh, a few internal links.
Rough equivalent ad spend based on current click prices. Directional, not an invoice.
Keywords where a competitor's specific page outranks yours, sorted by what the gap is worth. Each row is one page to build or upgrade.
NO PAGE = you have nothing ranking for that keyword at all
Over 65 characters. Google truncates titles past ~60-65 chars.
Under 30 characters. Likely missing keyword context.
Multiple pages share the same title. Confuses SERP intent.
Under 120 characters. Wasting SERP real estate.
No structured data found. Missing rich-result eligibility (Article, FAQ, Product, etc.).
Based on the 80 pages that returned HTTP 200 out of 80 fetched. Not a full-site crawl.
Click depth and inlink counts are computed over the 80 pages we fetched, not a full crawl. Pages outside the sample are not represented in the graph.
Compared the first 80 sitemap URLs against the top-500 ranked-keyword URL set.
Classic rankings and AI citations are now separate games. This checks whether the site has pages and entity signals worth retrieving.
Diginow is effectively invisible to AI answer engines. Zero AI Overview appearances against 100 each for morson-edge.com and cv-library.co.uk, and none of the 80 fetched pages match a citation-friendly format (best-of, alternatives, comparison, pricing explainer). The llms.txt manifest is present, which is more than most agencies bother with, but a manifest doesn't create anything worth citing. Schema covers 70 of 80 pages, so hygiene isn't the bottleneck. The gaps are answer-shaped pages and a clear entity: no Wikipedia entity was found, and the homepage H1 says "New Home page".
Zero of 80 fetched URLs match best-of, alternatives, comparison, review or pricing patterns. Nothing on the site is shaped the way answer engines quote.
Title and meta state category, location and services clearly, but the H1 is the placeholder "New Home page", so the strongest on-page entity signal is blank.
Case studies name real clients (Elemis, Headspace, CFP Energy) but no page publishes extractable numbers such as build cost bands, timelines or measured results.
No video or off-site entity signal appears in the data and no Wikipedia entity was found. llms.txt is present, which helps crawling but adds no third-party corroboration.
0 AI Overview appearances for diginow.co.uk against 100 for morson-edge.com, with 10 competitor-only opportunities recorded.
No AI-answer tracking in place. Citation sources shuffle week to week, so an unmeasured baseline means you can't tell whether any of this moved.
The formats AI answers cite most are best-of lists, comparisons and pricing explainers, and you have none. Start with "Best Web Design Agency in London" (320/mo), "eCommerce Website Developers in the UK" (390/mo) and a website design cost breakdown (590/mo, KD 0). Each needs a table of options with named criteria, not prose.
"New Home page" is the H1 on the page an answer engine is most likely to read when resolving what Diginow is. Replace it with a sentence naming category, service lines and both cities. Then give each /insights page its own H1 instead of "Pinboard & findings", so the AI SEO, recruitment and eCommerce articles read as distinct entities.
Answer engines quote specifics. Publish cost bands for bespoke CMS, recruitment and eCommerce builds, typical project timelines, and the measured outcome from each case study. Numbers in plain text under a matching heading get lifted; adjectives don't.
Highest-leverage first. Keywords with a ready article on top.
Authority comes before content here. The opportunities below won't rank without link signals first.
Start with general business directories that fit your niche. browse our directory of 100+ free sources.
One keyword in, one published article out, every day. We'll ship against the 30-keyword list above.
REAL DATA ABOVE · 8 QUICK WINS, 12 ARTICLE IDEAS, 30 KEYWORDS FOR DIGINOW