Second Site · Content intelligence platform
Don’t just see what’s wrong. See what’s next.
One platform that connects your website data, finds what needs fixing, improving and creating, and puts it in order of what matters to your business. Less time chasing broken links and missing alt text. More time on the work only your team can do.
Book a demo See how it worksWHY SECOND SITE?
Your tools are good at finding problems. And that's the problem.
Keeping a professional services website in good shape used to mean juggling five or six separate tools: an analytics platform, an SEO crawler, an accessibility checker and a content research tool. Every one of those tools has hundreds of features, and most of the time you use only a handful. Joining up what they tell you, every month, was a job in itself, and it fell on you. We've been in your seat, and we know that the setup fails you in three key ways:
Introducing Second Site:
One platform that connects the data, proposes the fixes, and puts them in order of what matters to your organisation. It takes the routine work off your team's plate, so the time goes on the interesting, creative work that got you into marketing in the first place.
How does it work?
See it. Improve it. Prove it.
Five layers, one continuous loop. Second Site starts by building a complete picture of your content and how it's performing, a ladder of agents then improve what they find based in your content.
Jobs
Journeys
ExpertisePrioritised for your business
Start with a clear picture of what you've got and how it's performing
Second Site starts by building a library of every page on your site. Every page is tagged to the service it sells, the sector it serves, the audience it helps and the questions it answers. Then we connect the sources that tell you how each page is performing: Google Analytics for who arrives and what they do, Search Console for where you show up in Google, your server logs for which AI platforms are reading which pages, and heatmaps for how people behave once they're there.
A crawler can tell you what's on a page. It can't tell you who came to it, what they did, or which AI platforms have been reading it. With the content and the performance data in one place, you can see how every page is doing today, and our Fix, Optimise and Create agents have something to work from.
Three layers of agents, each one doing more than the last
On top of your content library sit three layers of agents. Each layer works from the evidence on the page and from your context, checks it against a defined method, and recommends the specific changes most likely to improve performance.
- Fix tests every page on your site against hard and fast rules: technical health best practice and your target WCAG accessibility level. Where a link is broken or an image has no alt text, it tells you, and it separates the fixes your editors can make from the ones that need developer input.
- Optimise moves from fixing what's broken to improving what's already there. It looks at where a page could carry schema beyond what's already in the template, an FAQ block, how-to steps or a quotation, so that the content is machine-readable as well as human-readable. It also assesses how well each page demonstrates experience, expertise, authority and trust (E-E-A-T) for its subject, in line with Google's guidance. Each recommendation is specific to the page: what to add, where, and why it will help.
- Create looks beyond what's on the site today. It finds pages that are losing traffic and need refreshing, pages that answer the same question and get in each other's way, and gaps where a new piece should exist. Each opportunity becomes a data-driven brief your team can work from, written in the context of your audiences and the jobs they're trying to get done.
Devised and prioritised for your audiences, your services and your strategy
Off-the-shelf tools give everyone the same recommendations, weighted the same way. Second Site runs every layer against your context: who your audiences are, what each of them is trying to get done, and the expertise you're selling to help them do it. You set the priority on each, so the work tied to the audiences and services that matter most rises to the top, and if your priorities change, the recommendations reflect it.
Because of this context, when Second Site suggests a piece of content, it already knows who it's for, where they are in their decision, and what they need to hear from you. And because your tone of voice and brand guidelines are built into the same context, anything it drafts sounds like you.
From "good idea" to done, without the jumping around
Most tools stop at the list. Second Site carries the work through. Every content opportunity comes with a brief, and once you've approved it, the platform can write a first draft: structured to answer the questions the piece exists to answer, in your tone of voice, with your brand guidelines applied. Think of it as the first 80 per cent done for you. The 20 per cent that makes the piece yours, the insight only your people have, is what your team adds on top. Edit it, send it back for another pass, or approve it; every version is kept.
The next step takes the approved work into your site. A recommendation to add an FAQ block goes into the article as a draft. A new article goes into the CMS as a draft. A broken link gets fixed. Nothing is published without you; it lands in draft, so you never open the site and find changes live that you haven't seen.
Closing the loop nobody else can close
Every recommendation Second Site makes is logged, along with the date it goes live. From that point the platform tracks what the change did: to search rankings, to engagement, to conversions, and to how often the AI platforms read the page. Each crawl compares against the last, closes what has been fixed and flags anything that has come back. Over time you learn which changes move the numbers for your site and your audiences, and each round of recommendations is better informed than the one before.
That loop matters most where you are currently blind. Nobody can tell you how often an AI assistant mentions you in an answer, and your analytics cannot see the machines reading your site at all. Your server logs can. Because Second Site knows what changed and when, and reads the logs alongside it, it can show you how AI reads of a page rise or fall after a change is made. Nobody can promise that a given change will get you cited more often. This is the closest anyone can get to measuring it.
Why is it different?
Content built around the journey, not the keyword
Generating content is easy. Generating content worth reading is not. Most tools start from keywords or topics and work forwards, so what comes out is what every other organisation in your market is already publishing, only faster.
Second Site starts from the people you serve and works back: who they are, what they're trying to get done, and the journey they take to get there. Every recommendation and every draft exists because a real audience, at a real point in that journey, needed something your site didn't give them. Every piece has a reader before it has a title.
- Do I need a patent, or can I keep it as a trade secret?
- What's the difference between a patent and a trademark?
- A competitor has copied us. What can I do right now?
- How much does a patent application cost, start to finish?
- Can I file it myself, or do I need an attorney?
- How long does protection take, and what am I exposed to in the meantime?
- Do I need protection outside the UK?
- Who has done this for a business like mine?
- What will they need from me to get started?
- What does the first month look like?
Who it’s for
One platform, two reasons to look at it
If you run marketing:
Chief Marketing Officer
You need the website to work for the audiences that matter, and to show the board that it does. Second Site runs the whole channel against your strategy, so every fix, improvement and new piece is prioritised by your audiences and your services, and every change is recorded along with what it improved. It is the measurement behind a content strategy, not another dashboard.
If you run the website:
HEAD OF DIGITAL
You have a long list, a small team and a few hours a week. Second Site puts the analytics, the crawl, the accessibility checks and the content research in one place, turns the findings into recommendations rather than problems, and takes each one through to a draft you can review. The routine work stops eating the week, and the hours go on the work only your team can do.
How do we work with you?
A subscription that starts with a conversation, not a login
Second Site is a subscription, but it doesn't start with a login and a credit card. It starts with our team talking to yours, to understand the audiences you serve, the jobs they bring to your site and the services you want to be known for.
Your context is what every recommendation is prioritised against, and if it's thin, the recommendations will be too. So we don't leave it to a setup wizard. We do it with you, properly, once. You learn the method as we go, so adding an audience or changing a priority later is yours to do, whenever you need to.
If we don't think it's right for you, we'll tell you.
Second Site only earns its keep where there's a real audience to serve, content worth improving and someone on your side with the autonomy to act on what it finds. If we don't think it will make a difference to your business, we'll tell you.
What’s inside
Need some more detail?
Everything in the platform, layer by layer, for anyone who wants the detail:
1 · The core
A full inventory of every page on your site, discovered from your sitemap and rendered in a real browser where a page depends on JavaScript. Every page is tagged to the service it sells, the sector it serves, the audience it helps and the questions it answers, using rules you control. Pages that drop out of the sitemap are marked as removed, not deleted, and stop counting against your scores.
- Live performance per page: views, engagement, key events, clicks, impressions, average position
- Heatmaps and scroll depth alongside the search and engagement numbers, so you can see how people behave once they arrive
- Filter by service, sector, audience and content type, in any combination, with a date-range picker
- Bulk tagging, sortable columns, a pinned totals row and CSV export of exactly what’s filtered
- Multilingual sites: pages classified by language, with a language filter across every agent
2 · The agents
Twelve checks on every page: missing or duplicate meta titles, descriptions and H1s, title and description length, thin pages, broken internal links, redirect chains and orphaned pages. A site health score graded by severity, so one minor issue on a page is not treated as a failure.
- Drill into any issue type for the affected pages, how to fix it, and its trend over time
- Mark fixed, dismiss with a reason, assign or reopen, per page or in bulk
- Rescan a single page and see what was fixed and what is new, without waiting for a full crawl
- View on page: a screenshot with the problem highlighted
Auditing against your target WCAG level, run with the industry-standard axe-core engine on each fully rendered page. Every issue is classed as editorial (your content team can fix it) or technical (a developer fixes it in code), with a plain-English description of the affected element, the WCAG criterion, the HTML and a suggested fix.
- A score that counts only issues at your target level, plus a view of who is affected by disability category
- An elements view that groups the same fault across many pages, so a template fix is assigned once
- Items that need human judgement are flagged for review rather than guessed
- Cookie banners dismissed before testing, so they do not hide the page
Schema in three layers: site-wide identity (set up once), the markup each page type should carry (set up once per template), and the in-body markup editors add page by page, which is the ongoing work. Second Site reads each page and flags where an FAQ, a how-to, a quotation, a video, an image or a review could carry schema, and which CMS component to use.
- Layer status per page: complete, partial, missing or malformed
- Evidence for each opportunity, where on the page it applies, and the expected impact
- Free per-page rescan, so a newly added component is picked up straight away
Every page scored from 0 to 100 on experience, expertise, authoritativeness and trust, in line with Google’s guidance, against the right model for the page: service, article or people. The signals it looks for run from named clients and case examples to credentials, citations of legislation and the courts, publication dates, regulator registration and author schema.
- Evidence behind every score, the page’s strengths, and up to six improvements ranked by impact
- Create a brief from any improvement, straight into Content Studio
- The signal vocabulary is tuned for your organisation: the credentials, citations and regulators that count in your sector
What to create, refresh or consolidate, and why. Create: a question your audience asks with no page answering it, or search demand you rank weakly for. Refresh: a page that used to perform and is losing traffic. Consolidate: two or more pages answering the same question and getting in each other’s way.
- Every opportunity is tied to a specific audience, job, stage and page, and carries that job’s priority
- Each opportunity carries a brief seed: structure, keywords, target length, internal links and a schema recommendation
- A work queue with the same statuses and actions as every other agent, reconciled on every scan
3 · The context
The structure that makes the recommendations yours. Expertise groups and areas, audience groups and audiences, each with a priority you set. Jobs link audiences to expertise, and each job carries a journey: the questions people ask at the trigger, comparison and decision stages. Your tone of voice and brand guidelines sit alongside, and everything the platform writes uses them.
- Priorities: Tier 1, 2, 3 or inactive on every audience and expertise area, so the work that matters most rises to the top on every screen
- Questions suggested by the platform from your content and your search data; you confirm or reject every one, and add your own
- Every crawl works out which confirmed questions each page answers, with a word-for-word quote from the page as evidence
- A coverage view for every job: which questions your site answers fully, partly or not at all
4 · Deliver
From brief to publish-ready draft, in your voice. Three stages: briefs, in review, ready to publish. Start a brief from a content opportunity, from an E-E-A-T improvement, from a topic with the platform’s help, or by hand.
- A brief holds the audience, the angle, the questions the piece must answer, the section structure, keywords, target length, internal links and a schema recommendation
- Generate a full draft in around 15 to 20 seconds, in UK English, using your tone of voice and brand guidelines
- Edit in the browser; every save is a version; regenerate with editor notes
- Approved work goes into your CMS as a draft, for you to publish
5 · Measurement
Every recommendation is logged, along with the date it goes live, and each crawl compares against the last: fixed issues close, new ones open, and anything that comes back is flagged as a regression. Where your server logs are connected, the platform also shows which AI platforms are reading which pages: training crawlers, search indexers, and live fetches to answer a real question. Your analytics cannot see any of that.
- Per-page trends for search, engagement and AI reads, before and after each change
- AI bot hits per page, per day and per operator, with identity checks so impersonators are kept separate
- Visitors who arrive from an AI answer, which analytics attributes to “direct” or loses altogether
- Dismissed items stay dismissed across scans; nothing you have ruled out comes back to nag you
Experience for users
Everything about one page in one place: performance, search queries with change over time, AI reads, technical issues, accessibility issues, structured data status, E-E-A-T scores and every content opportunity involving the page. Six tabs, each linkable directly. Rescan any one of them without waiting for a full crawl.
The same statuses everywhere: open, needs review, regression, fixed, dismissed. The same actions everywhere: mark fixed, assign, dismiss with a reason, reopen. Bulk actions on many items at once.
- Assign any issue, finding or opportunity to a named colleague, with a note
- My Work: a personal inbox of everything assigned to you and by you, grouped by agent
- Every status change recorded: who, when and why, including changes the system made
- CSV export from the library and every major agent
Book a demo
See it running on your own site
A 45-minute walkthrough, using your site as the example. We’ll crawl it beforehand, so you see real findings rather than a demo account.
Book a demoNot ready for a demo?
Start with the Perception Gap Scorecard: three minutes, and an indication of where your site’s biggest gap is. Or ask for a free Second Opinion, and see how your site reads to one audience of your choice and to the machines.
FAQs
Questions we often get asked before a demo
Yes. Second Site crawls your site from its sitemap and audits the pages it finds, so it works with any CMS. It renders JavaScript-heavy pages in a real browser and can scan a staging site behind a login. Getting content back into the CMS is copy and paste today; direct publishing is on the way, starting with sites we host.
Three ways. It’s continuous rather than a snapshot: every crawl compares against the last, closes what’s fixed and flags what’s come back. It prioritises against your strategy rather than a generic score: your audiences, jobs, questions and expertise, tiered and confirmed by you. And it does the work rather than listing it: from finding to brief to draft, with your sign-off at every step.
No. Second Site finds, recommends and drafts. A named person on your team reviews and approves before anything reaches your site, and every change is logged with who made it and when.
Your sitemap, read-only access to Google Analytics 4 and Search Console through Google sign-in, and a working session on your audiences, the jobs they come to do and the services you want to be known for. If you’ve already done audience mapping with us, that work feeds straight in. If you haven’t, it’s where we’d start.
From your server logs, not from asking chatbots. Prompt-based scores can’t be reproduced: the same question gives a different answer each time. What we can count is what the AI platforms read on your site, whether they fetch a page live to answer a real question, and the visitors who arrive from an AI answer. That comes from integrating your log files, and it’s the one thing your own analytics can’t show you.
Most of the platform doesn’t use AI at all: crawling, technical checks, accessibility testing, scoring and trends are deterministic. AI is used where judgement adds value, such as suggesting questions, linking pages to questions, writing up opportunities and drafting content. Where it’s used, its output is evidenced, checked against the page, and waits for you to confirm.
Yes. Access is invite-only, each organisation’s data is isolated at the database level, Google connections are read-only, and access tokens and any scanning credentials are encrypted at rest.
Crawling, classification, filtering and duplicate detection are language-aware, and you can scope the whole platform to one language. The AI features, including question suggestions and drafting, currently work in English.
If your website exists to help specific audiences get specific things done, yes. Second Site is built for professional services: law firms, consultancies, accountancy practices, building societies and other financial services. The audiences, jobs and questions structure is yours, and the skills the agents work from are tuned for your organisation, so the credentials, citations and regulators that signal expertise in your sector are the ones it looks for. It works on any CMS and on sites from a few hundred pages to several thousand.
Second Site is a subscription, priced by the size of the site and the agents you switch on. There's a one-off set up cost used to cover the audience mapping and context configuration. We’ll give you a figure on the demo call once we’ve seen the site.