LongTermSoftware.com Live Audit Report

Full report

On this page

Generated UTC: 2026-06-17T01:49:40Z Target: https://longtermsoftware.com/ Audit mode: Non-invasive public-site review. No WordPress login, no live modifications, and no real lead-data submission.

Scope and evidence note

This report follows the uploaded audit brief and uses the live production site as the source of truth. Evidence was gathered through the web reader, local downloaded public artifacts, PDF extraction/rendering, and search-result checks. Direct browser screenshots, raw curl -I status-code checks, Lighthouse, axe, DevTools network inspection, and keyboard testing could not be completed because the active local container could not resolve longtermsoftware.com with curl or headless Chromium. The failure is preserved in logs/browser-network-limitation.log and referenced below. Findings are separated accordingly.

Evidence reference map

---

1. Executive Summary

The live site is substantially aligned with the expected public truth. The homepage presents the company as an enterprise software modernization and reviewed-AI consultancy, the primary navigation appears in the required order, the pricing ladder is visible and internally consistent across the HTML page and machine-readable assets, and the contact details match the expected Mike Kappel / mike@ns12.com / (708) 230-2304 / Cicero, IL truth.

The strongest live areas are pricing truth, claim boundaries, machine-readable readiness, and public proof discipline. The site avoids unsupported AI claims and repeatedly distinguishes templates, generated artifacts, and approved outcomes. This is a credible posture for compliance-conscious buyers and AI-agent/vendor-evaluation tools.

The highest-priority issues are not primarily content accuracy issues. They are verification and buyer-flow issues:

1. Add real browser/CI evidence for the required screenshots, viewports, Lighthouse, axe, keyboard, and diagnostic network tests.
2. Clean stale search-index results for Hello world!, /projects/, and old resume/recruiter /downloads/ snippets.
3. Reduce overlap between Resources, Downloads, and Docs.
4. Make first-touch contact and fit-call flow shorter and more accessible.
5. Add diagnostic copy/export and a runtime privacy guard test.

---

Required Table A. Overall scorecard

---

2. Current Live State vs Expected Site Truth

---

3. Screenshot Catalog

No PNG screenshots were captured in this environment. The web reader provided live text evidence, but the screenshot tool available here only supports PDFs, and local Chromium could not resolve the production domain. The next theme-build QA run must capture the following files with a desktop browser or Playwright CI.

---

4. Route and Link QA

Required Table B. Route audit table

Link and route observations

• No confirmed broken internal link was found from the web-reader extractions.
• Confirmed redirect behavior exists for /hello-world/ to the homepage and /projects/ to /proof/, but search results still expose stale URLs/snippets.
• Many old aliases and utility files need CI HEAD/GET verification because direct local browser/curl checks were unavailable.

---

5. Homepage and First Impression Review

Confirmed live strengths

• The hero explains the company quickly: modernization and AI for systems that cannot tolerate behavioral drift.
• The technology fit is explicit: .NET, SQL Server, Web Forms, Classic ASP, VB.NET, and stored-procedure-heavy systems.
• Above-the-fold trust and buyer-path signals are present: ad-hoc entry price, monthly advisory price, and direct contact.
• Unsupported claim boundaries are visible later on the page, including no AGI/consciousness/safety certification/vendor partnership/autonomous production authority claims.

First-impression risk

The first screen is now sales-oriented enough to be credible. The risk is that lower sections still behave like a technical proof library rather than a conversion path. That is not a fatal problem for this brand, but it should be controlled: buyers should move from problem -> proof -> service -> fit call, not problem -> machine-readable proof archive.

Proposed homepage section order

1. Hero: pain + fit + CTA + direct proof markers.
2. Buyer problem strip: legacy drift, stored procedures, unsafe AI, undocumented rules.
3. Entry paths: Ad-hoc, Monthly Advisory, Annual Modernization, limited C2C.
4. Service chooser: seven service cards with one-line fit and next CTA.
5. Proof ledger preview: 3-4 proof cards, each marked template/outcome/status.
6. Browser-local diagnostic: privacy-first triage tool.
7. Contact/fit call: short form + direct phone/email.
8. Machine-readable evidence: subdued footer-level technical proof.

---

6. Header, Navigation, and Search Review

Confirmed

• Header navigation text appears as Services, Proof, Industries, Resources, About, Contact.
• Search appears with text equivalent Search LongTermSoftware.com Search.
• Fit call is present in the header.
• A skip link appears before header content in extracted pages.

Not verified

• Header wrapping at 1366, 1280, 1024, 768, 390, 375, and 320 CSS px.
• Mobile menu open/close behavior.
• Keyboard focus order and Escape behavior.
• Tap-target sizing and focus-visible contrast.

Recommended CSS/header fixes

• Use a container-query or breakpoint where search collapses before nav wraps.
• Keep the fit-call CTA visible until the mobile menu state, then keep it as the first action inside the opened menu.
• Add aria-label="Primary navigation", a labeled search input, and an aria-expanded menu button.
• Set a route QA assertion: document.documentElement.scrollWidth <= window.innerWidth for all required widths.

---

7. Pricing Page Review

Required Table D. Pricing consistency table

Assessment

The pricing page is one of the strongest parts of the site. The four-tier ladder is clear, and the non-accumulative retainer language is explicit enough for procurement and AI readers. Monthly Advisory is correctly framed as the practical continuity/default option, Annual Modernization as protected calendar capacity, and Embedded Architect C2C as limited availability rather than the default.

Implementation refinements

• On mobile, keep each card’s tier name, rate, commitment, and CTA visible without excessive scrolling.
• Use consistent CTA labels:
• Ad-hoc: Request focused architecture help
• Monthly: Discuss monthly advisory fit
• Annual: Discuss modernization capacity
• C2C: Check embedded architect availability
• Keep the aria-label more descriptive than the visible CTA where shortening is needed.

---

8. Contact and Fit-Call Review

Required Table E. Contact consistency table

Confirmed

• Contact and fit-call pages show Mike Kappel, mike@ns12.com, (708) 230-2304, and Cicero, IL.
• Phone/email actions are present as visible contact options.
• Privacy warning is visible: the site tells users not to paste secrets, private source code, PHI, credentials, or confidential data.

UX issues

• The contact page includes a detailed assessment intake. This is useful but likely too heavy as a default first-touch path.
• Form label/error behavior could not be verified.
• Confirmation path could not be safely tested without submitting data.

Field-level recommendations

---

9. Services UX Review

Recommended service-card structure

1. Service name.
2. Buyer problem in one sentence.
3. Best fit.
4. Timeline.
5. Pricing path.
6. Deliverables.
7. Proof/sample link with explicit status.
8. CTA: Discuss this path.

---

10. Proof, Case Studies, and Sample Artifacts Review

Proof maturity assessment

The proof system is more mature than typical small consultancy sites. It has a claim-to-evidence ledger, machine-readable assets, status labels, route QA, and explicit unsupported-claim boundaries. That is a differentiator for procurement, technical reviewers, and AI vendor-evaluation tools.

Template-vs-outcome boundary

The boundary is clear: sample artifacts are marked as templates, generated artifacts require human review, and approved client outcomes require approval. This should remain non-negotiable in the next theme package.

Top 5 proof improvements

1. Add a buyer relevance line to each proof card.
2. Add template, generated draft, approved public outcome, or machine-readable evidence badges using buyer-friendly text.
3. Add a short why this matters paragraph to each case-study path.
4. Add one approved outcome or testimonial only after human/client approval.
5. Add proof-to-service cross-links so each service card has specific evidence.

Recommended proof-card format

---

11. Resources and Downloads Review

Resource inventory

Observations

• Live Resources and Downloads pages are buyer-facing and useful.
• The current overlap between Resources, Downloads, and Docs creates taxonomy friction.
• Search snippets still suggest old recruiter/resume content on /downloads/, despite live page content being buyer-facing.

Recommended resource taxonomy

---

12. Browser-Local Diagnostic Review

Confirmed from live page text

• It says it runs as local keyword matching.
• It says no data leaves the browser.
• It says no model call is made.
• It warns users not to paste secrets, credentials, source code, PHI, or confidential data.
• It includes example/sample diagnostic text and maps to next steps.

Not verified

• DevTools Network evidence after entering the required safe sample input.
• Whether typed input appears in request bodies, query strings, local storage, analytics, or logs.
• Whether result copy/export is present and usable.
• Whether the result is announced to screen readers.

Recommendations

1. Add a Copy sanitized result button after local analysis.
2. Add a Use this in fit-call request CTA, but do not auto-submit or transmit diagnostic text.
3. Add a route QA network test:

• Visit /diagnose-system/.
• Enter the safe sample input.
• Click run.
• Assert no POST occurs.
• Assert no request URL/body/header contains diagnostic text fragments.

1. Add aria-live="polite" to the result region.
2. Track diagnostic_start, diagnostic_complete_local, and diagnostic_result_cta_click without raw input.

---

13. Machine-Readable UX and AI-Agent Readiness

Exact mismatches

• No content contradiction was confirmed between pricing HTML, pricing-ladder.json, llms.txt, and AI manifest.
• The main mismatch risk is version perception: proof-ledger JSON uses 1.7.0 while the site/manifest/route QA assets use 1.18.0. That may be valid if it is a data/schema version, but the public files should make that distinction explicit.
• Several static utility files and REST endpoints remain unverified in this run and must be tested by CI.

---

14. Accessibility Review

Required Table F. Accessibility table

Automated findings

No automated axe/Lighthouse/WAVE result is available from this run because local browser access failed. Do not treat the accessibility score as complete.

Manual findings

Manual text extraction suggests good baseline structure: skip links, clear H1s, descriptive CTAs, descriptive download links, and explicit privacy/boundary text. The main remaining risks are interactive: menu, search, form errors, diagnostic result update, focus state, contrast, touch targets, and reflow.

---

15. Mobile and Responsive UX Review

Required Table C. Viewport table

_The full viewport matrix is included in viewport-audit.csv inside the artifact package._

Required acceptance checks for next run

• Header does not wrap at 1366, 1280, 1024, 768.
• Logo remains visible and proportional.
• Search collapses or reflows before creating overflow.
• CTA remains visible or appears as first item in mobile menu.
• Cards remain readable at 390, 375, and 320 widths.
• No horizontal scroll at 320 CSS px.
• Forms remain usable with zoom and text resize.

---

16. SEO and Index Hygiene

SEO defect table

Recommended title/meta rewrites

---

17. Performance and Technical UX

Performance summary

Lighthouse and page-weight metrics are unavailable from this run. Do not infer performance scores from text extraction.

Required next-run metrics

Likely optimization areas to inspect

• Hero and illustration image dimensions and compression.
• Render-blocking WordPress CSS/JS.
• Font loading and font-display behavior.
• Unused CSS from theme scaffolding.
• Lazy loading for below-the-fold illustrations.
• Cache headers for JSON, CSV, PDFs, images, CSS, and JS.

---

18. Security and Privacy

Quick-risk table

---

19. Content and Messaging

Strong copy examples

• Corporate software architecture for systems that cannot drift when AI or modernization changes the stack.
• No data leaves your browser. No model call is made from this page.
• Retainer hours must be used in the calendar month where they are reserved.
• Evaluation support is not safety certification.

Weak or risky copy examples

---

20. Severity-Ranked Findings

Required Table G. Severity findings table

---

21. Implementation Backlog

Required Table H. Implementation backlog

---

22. Exact Copy Recommendations

Additional microcopy:

• Diagnostic empty state: Paste a sanitized system summary to receive a local triage result. Do not include secrets, credentials, private source code, PHI, or confidential client data.
• Diagnostic result disclaimer: This is a triage aid, not production architecture advice or legal/compliance guidance.
• Contact form helper: A short summary is enough. Use sanitized details only.
• C2C card note: Limited availability; not the default first engagement.
• Proof badge labels:
• Template — not a client outcome
• Generated draft — human review required
• Approved public outcome
• Machine-readable evidence

---

23. Theme/File Change Recommendations

---

24. Tests to Add to Route QA

1. route_status_matrix: GET every required route, old alias, download, static file, and REST endpoint; record status, final URL, content type, canonical.
2. title_meta_canonical_matrix: parse <title>, meta description, canonical, robots meta, H1.
3. viewport_screenshot_matrix: capture screenshots at 1440, 1366, 1280, 1024, 768, 390, 375, and 320 widths.
4. no_horizontal_scroll: fail if scrollWidth > innerWidth at any required viewport.
5. header_contract: assert logo visible, nav order stable, CTA accessible, mobile menu works, search has accessible name.
6. pricing_truth_consistency: compare HTML, pricing JSON, REST pricing, pricing consistency, llms, AI manifest, FAQ.
7. contact_truth_consistency: compare contact page, fit-call page, footer, schema, llms, AI manifest, REST contact profile.
8. unsupported_claim_scan: fail on AGI/consciousness/safety certification/vendor partnership/guaranteed compliance/autonomous production authority/unreviewed outcomes.
9. diagnostic_network_privacy: enter safe sample; assert no user text sent in requests and no POST triggered.
10. diagnostic_a11y: result region uses aria-live, buttons keyboard-operable, copy result works.
11. axe_smoke: run axe on home, pricing, contact, fit-call, services, proof, resources, diagnostic.
12. lighthouse_ci: save mobile and desktop JSON/HTML reports for required routes.
13. pdf_accessibility_inventory: verify each PDF has HTML equivalent, title metadata, and a PDF accessibility status marker.
14. download_schema_validation: validate JSON/CSV columns, generated UTC, version fields, and content-type headers.
15. serp_hygiene_manual_check: generate a checklist for Search Console review of stale starter/recruiter snippets.

---

25. Final Recommendation

Proceed with the next WordPress theme package, but treat the next pass as a QA-hardening and conversion-clarity pass rather than a full positioning rewrite. The site’s public truth is mostly sound. The theme should preserve pricing/contact/proof boundaries and focus on:

• measurable browser evidence,
• stale SEO cleanup,
• shorter first-touch flow,
• stronger diagnostic conversion,
• cleaner resource taxonomy,
• accessibility and performance automation.

Do not add unsupported AI claims, invented client metrics, false partnerships, compliance guarantees, safety certifications, autonomous production authority, or unreviewed production outcome claims.

---

26. Prompt for the next AI theme-build pass

You are building the next WordPress theme ZIP for LongTermSoftware.com. Use the current live site as source truth and preserve the confirmed public positioning: .NET/SQL modernization, behavior-preserving legacy upgrades, Web Forms, Classic ASP, VB.NET, SQL Server/stored-procedure-heavy systems, human-reviewed AI workflows, source-bound AI documentation, governed knowledge/RAG foundations, AI reviewer applications, AI evaluation/reliability, and systems that cannot tolerate behavioral drift or unsupported AI claims.

Top implementation priorities:
1. Add Playwright route QA for all required desktop/tablet/mobile viewports: 1440x900, 1366x768, 1280x800, 1024x768, 768x1024, 390x844, 375x667, 320x568.
2. Generate named screenshots for home, header, pricing, contact, fit-call, services, proof, resources, diagnostic before/after, and machine-readable evidence.
3. Add Lighthouse CI for home mobile/desktop, pricing mobile, contact mobile, and diagnostic mobile. Save JSON/HTML reports.
4. Add axe-core and keyboard tests for skip link, nav/menu, search, CTA focus states, forms, diagnostic, and reflow.
5. Clean SEO/index hygiene: remove starter-content /hello-world/ exposure, stale /projects/ index, and old resume/recruiter /downloads/ snippets.
6. Refactor /docs/ into a distinct technical evidence/public memory hub, not a Resources duplicate.
7. Preserve pricing truth exactly: $150/hour ad-hoc no minimum; $100/hour monthly advisory, 20-hour monthly minimum, $2,000/month; $80/hour annual modernization, 40-hour monthly minimum, $3,200/month, annual contract; $75/hour embedded architect C2C, 40 hours/week, six-month minimum, limited availability.
8. Preserve non-accumulative retainer language: unused hours do not roll over, accrue, bank, or carry forward unless a signed order form says otherwise.
9. Preserve contact truth exactly: Mike Kappel / Michael Kappel, mike@ns12.com, (708) 230-2304, Cicero, IL.
10. Improve diagnostic conversion: local-only network guard, copy-result button, result CTA, aria-live result region, and analytics events that never capture pasted text.

Likely files to change: header.php, footer.php, front-page.php, templates/virtual-page.php, inc/forms.php, inc/schema.php, inc/rest.php, inc/content.php, inc/enhancements-v*.php, assets/css/app.css, assets/js/app.js, route-qa-contract.json, tools/route-qa.py, llms.txt, llms-full.txt, ai-agent-manifest.json, .well-known/ai-plugin.json, .well-known/security.txt, .uai/*, docs/*.

Claims not to make: AGI, consciousness, safety certification, formal vendor partnership, guaranteed compliance, autonomous production authority, unreviewed production outcomes, unsupported client metrics, automatic repository writes, hosted UAIX importer behavior, or runtime authority from machine-readable files.

Acceptance criteria:
- Every required route returns expected status/final URL/content type/canonical/title/H1/markers.
- All pricing/contact sources match HTML, REST, llms, AI manifest, and downloadable JSON/PDF/CSV assets.
- Header does not wrap or create horizontal scroll at required widths.
- Mobile menu, search, forms, CTAs, and diagnostic are keyboard-operable and accessible.
- Lighthouse/axe artifacts are saved.
- Diagnostic sample input is never sent over the network.
- Search-index cleanup plan is documented for Search Console.
- Public proof/sample artifacts clearly distinguish templates, generated drafts requiring human review, machine-readable evidence, and approved client outcomes.

Final response requirements for the theme-build agent:
- Provide the theme ZIP.
- Provide route-qa output, screenshot catalog, Lighthouse reports, axe findings, and exact changed files.
- State any items requiring WordPress admin, analytics/Search Console/server access, or human approval.