
Posts
Posts


A free 40-point scorecard for your careers website: 8 areas, 5 checks each, clear criteria for Yes, Partly and No, and a benchmark from 60 large employers so you can see how your results compare.
A careers website is where candidates decide whether to apply. Problems usually sit in different teams: speed with IT, search visibility with marketing, content with employer brand, and the application with talent acquisition. A scorecard puts everything on one page, so you can agree what to fix first. Most checklists for this topic are generic website audits. This one is built for careers websites and for the people who own them.
The bands are our judgement of what the totals mean. They are not drawn from the benchmark data.
Each area has five checks. A "measured" check can be tested with free tools; an "expert criterion" is our judgement of good practice.
The full scoring criteria for every check (what counts as Yes, Partly and No, and how to test it) are in Appendix A and in the spreadsheet.
Share scoring Yes, higher is better. S = speed, E = search visibility, A = accessibility, J = job pages.
Lab tests on one home page and one job page per site. The speed results are low partly because the test is strict.
The chart shows the share of the 60 sites that earned full marks on each of the 14 measurable checks. The speed results are low partly because the test is strict: Lighthouse simulates a mid-range phone on a throttled mobile network, so use the result to compare your pages over time and against others tested the same way, rather than as a verdict on real-world experience. Search visibility checks show where the gains are: only 18% of sites carried a validThrough date, and 47% had the job title, text and markup all in the initial HTML.
Across all 14 checks the median site scored 6.8 (range 3 to 10.5). That means a typical large employer has room to improve on roughly half of the measurable points.
Out of 14, higher is better. The subgroup is reported separately and is not part of the 60-site sample.
Happydance subgroup (8)Other sites (60)
Indicative only: 8 sites, automated checks.
The 8 Happydance careers websites in our test scored a median of 8.5 out of 14 on the same checks, against 6.8 for the 60-site sample. The subgroup is reported separately and is not part of the sample. With 8 sites it is indicative only, and the checks are automated.
Compress and resize images and video posters, defer third-party scripts (chat, analytics tags, consent tools) and remove unused code. Re-run the test on a mobile profile after each change.
Add complete JobPosting markup in the page HTML (not injected later), add validThrough and remove or mark expired jobs, make sure canonicals and the sitemap are correct, and keep titles short.
Fix colour contrast and tap target size in the design system first, since they repeat across pages. Then run the keyboard, zoom and form tests by hand. See the careers website accessibility guide.
Test the journey on a real phone, simplify the first screen to one main action, and add a route for candidates with no suitable job today (alerts or a talent community).
Write the Give and Get in specific words, refresh employee stories with real people, and add a page for each big job family. Use evidence candidates can check.
Standardise a job page template: plain title, short summary, scannable sections, pay where permitted, a visible apply button and links to similar jobs.
Apply for your own job on a phone and time it. Remove fields you do not use, allow saving, remove the account wall where you can, and confirm what happens next.
Name an owner, tag your links, and set up a monthly view of visits, job views, application starts and completions. Put the quarterly re-check in the calendar.
Related reading:
A careers website audit should cover speed, search visibility (including job posting structured data), accessibility, design on mobile, employer content, job page quality, the application journey, and how you measure and own the site. The scorecard covers these in 40 checks.
You test a small set of real pages (the home page, a job page and the application), score each check against clear criteria, add up the points by area and agree the biggest fixes. The scorecard turns that into a score out of 40.
The scorecard takes about 90 minutes for one person with a laptop and a phone, by our estimate. A deeper audit of every page template takes longer.
At least quarterly for the measured checks, because speed, search visibility and accessibility can change with every release. Review the content and application checks twice a year.
Yes. The spreadsheet and the printable checklist are free to use.
Yes. The checks are vendor-neutral and apply whatever your site is built on, including an applicant tracking system career page.
Ideally a small group: someone from talent acquisition or employer brand who owns the site, plus colleagues from marketing, IT or web, and an accessibility lead if you have one. One person can score it, but fixes need the group.
If you own a careers website that was tested and believe a result is wrong, tell us and we will check it and correct it. Employer names are not published, and results are shown only in aggregate.
| # | Check | Yes if (1) | Partly if (0.5) | No if (0) |
|---|---|---|---|---|
| S1 | Home page Largest Contentful Paint (LCP) on a mobile test | LCP 2.5 s or less | LCP over 2.5 s and up to 4.0 s | LCP over 4.0 s |
| S2 | Job page LCP on a mobile test | LCP 2.5 s or less | LCP over 2.5 s and up to 4.0 s | LCP over 4.0 s |
| S3 | Lighthouse mobile performance score on both pages | 90 or more on both pages | 50 or more on both pages | Under 50 on either page |
| S4 | Total Blocking Time (TBT) on both pages, a lab measure of how long the page is busy | 200 ms or less on both pages | 600 ms or less on both pages | Over 600 ms on either page |
| S5 | Cumulative Layout Shift (CLS) on both pages | 0.1 or less on both pages | 0.25 or less on both pages | Over 0.25 on either page |
| E1 | Job pages carry Google JobPosting structured data with every required property | JobPosting present with all required properties | JobPosting present but a required property is missing | No JobPosting markup |
| E2 | Job title, text and markup are in the initial HTML, not added only by JavaScript | All three are in the source | Some but not all are in the source | None are in the source |
| E3 | Jobs carry a validThrough date, or expired jobs are removed or return a clear not-found status | validThrough is present (or expired jobs are reliably removed) | Handled for some jobs only | Expired jobs stay live with no expiry signal |
| E4 | Recommended properties: employmentType, identifier, directApply and baseSalary where you can state pay | 3 or 4 of the 4 are present | 2 of the 4 are present | 0 or 1 of the 4 are present |
| E5 | Job pages are not noindexed or blocked, have a canonical that points to themselves, and a sitemap is declared in robots.txt | All four are true | Three are true | Two or fewer are true |
| A1 | Lighthouse accessibility score on both pages | 90 or more on both pages | 70 or more on both pages | Under 70 on either page |
| A2 | No colour contrast failures on the home page or a job page | No failures on either page | Failures on one page | Failures on both pages |
| A3 | Buttons and links are big enough to tap, at least 24 by 24 CSS pixels or with enough spacing | No failures on either page | Failures on one page | Failures on both pages |
| A4 | A keyboard user can search, open a job and complete the application, and always sees where focus is | Everything works and focus is always visible | Works with minor problems (for example weak focus styling) | Blocked, or focus is lost or trapped |
| A5 | Pages stay usable at 200% zoom and at 320 CSS pixels wide, with no sideways scrolling or lost content | Fully usable at 200% and 400% | Usable at 200% with some problems at 400% | Content is lost, overlaps, or needs two-way scrolling |
| D1 | Searching, reading and applying all work comfortably on a phone, without pinch zooming | Smooth from search to submit | Works but with friction in one step | A step is hard or impossible on a phone |
| D2 | The home page shows one clear primary action (for example search jobs) in the first screen, not covered by banners or pop-ups | One clear action, nothing covers it | Action is present but competes with banners or other prompts | No clear action, or a banner or pop-up blocks it |
| D3 | The careers website looks and sounds like your brand and links smoothly to the corporate site | Consistent look, tone and navigation | Mostly consistent, with a jarring step | Looks like a different organisation |
| D4 | Candidates can search and filter by the things they care about (for example location and job family), and no-result searches help them | Search and filters work well and no-result pages offer alternatives | Search works but filters or no-result help are weak | Search is missing, slow or often returns nothing useful |
| D5 | Candidates with no suitable job today can leave details or sign up for alerts | Easy to find and works | Exists but is hard to find | Not available |
| C1 | The site says in plain, specific words what you offer people and what you expect from them | Specific, believable and easy to find | Present but generic | Missing |
| C2 | Stories from named employees of different roles, levels and places, kept up to date | Varied, real and updated in the last 12 months | Few, old or all from one group | None, or only stock content |
| C3 | Each main job family has a page that shows what the work is like and what the career path is | Pages exist for all of them | Pages exist for some | No such pages |
| C4 | Candidates can see where roles are based, and whether they are office-based, hybrid or remote | Clear for every location and role type | Clear for some | Missing |
| C5 | Benefits, development and (where you can) pay information is concrete rather than a list of buzzwords | Concrete, with examples or numbers | Mostly generic | Missing |
| J1 | Job pages have a title tag of 60 characters or fewer and a meta description | Both on all three jobs | Only one of the two | Neither |
| J2 | The job title is plain and searchable, and a short summary at the top says what, where and what type of role | Plain title and a clear summary | Title is clear, summary is missing or long | Internal job codes or unclear titles |
| J3 | The description uses short sections and lists (role, requirements, benefits), and avoids unnecessary jargon | Easy to scan and to the point | Mixed quality | Wall of text |
| J4 | Pay or salary range is shown where your market allows or requires it | Shown on jobs where it is allowed or required | Shown on some jobs | Not shown, although it could be |
| J5 | An apply button is visible without hunting, and the page links to similar jobs, the team or the location | Visible apply button and useful onward links | Apply is visible but there is nowhere else to go | Apply is hard to find |
| V1 | A candidate can complete and submit the application on a phone | Completes easily | Completes but with friction | Fails or is very hard |
| V2 | Candidates can start applying without first creating an account | Can apply without an account (or sign-in is optional) | Account is needed but quick to create | Long registration before applying |
| V3 | The application asks only for what is needed, shows progress and lets people save and return | Short, shows progress and can be saved | One of these is missing | Long, with no progress or save |
| V4 | Fields have labels, errors say what is wrong and how to fix it, and nobody re-enters information already given | All true | Mostly true | Labels or errors are missing |
| V5 | After applying, candidates see a confirmation and are told what happens next and when | Confirmation plus clear next steps and timings | Confirmation only | Nothing, or an error-like page |
| M1 | You can see visits, job views, application starts and completed applications | All four, reviewed regularly | Some are tracked | Not tracked |
| M2 | Job ads, social posts and emails use tagged links so you know which channels produce applications | Used consistently | Used on some channels | Not used |
| M3 | Someone owns the careers website, and content changes take days rather than months | Named owner and changes in days | Owner is clear but changes are slow | No clear owner |
| M4 | Speed, search visibility and accessibility are re-checked at least quarterly | Re-checked at least quarterly | Checked once a year or after a rebuild | Never |
| M5 | Applicant privacy notice, cookie choices and a way to report an accessibility problem are easy to find | All easy to find and accurate | Some are present | Missing |
| # | Check | Yes (of 60) | Partly | No |
|---|---|---|---|---|
| S1 | Home page loads fast on mobile | 0% | 3% | 97% |
| S2 | Job page loads fast on mobile | 0% | 5% | 95% |
| S3 | Lighthouse performance score | 0% | 10% | 90% |
| S4 | Page responds quickly to taps | 5% | 15% | 80% |
| S5 | Page does not jump around | 38% | 13% | 48% |
| E1 | Job posting markup is complete | 72% | 10% | 18% |
| E2 | Job content is in the first page load | 47% | 45% | 8% |
| E3 | Expired jobs are handled | 18% | 0% | 82% |
| E4 | Recommended job properties | 47% | 28% | 25% |
| E5 | Indexing basics | 37% | 48% | 15% |
| A1 | Automated accessibility score | 77% | 23% | 0% |
| A2 | Text contrast | 58% | 23% | 18% |
| A3 | Touch target size | 68% | 13% | 18% |
| J1 | Title tag and meta description | 32% | 65% | 3% |
The 40-point scorecard comes as a spreadsheet and a printable checklist, free to use.