7 Technical SEO Audit Steps 2026 [Pricing Checked]

A technical SEO audit for an online directory is a crawl-and-index pass over listing pages, category hubs, and city landing URLs that finds duplicates, pagination traps, and recrawl delays before Googlebot wastes the week on thin user-generated copies. It is not a sales-team lead score, it is not a review widget, and it is not a homepage copy refresh. Screaming Frog, Sitebulb, Lumar, OnCrawl, Botify, Semrush, and Ahrefs are the seven tools in this shortlist. US Tech Automations sits above them as the ticket layer that blocks a listing-template deploy when indexability, canonical, and duplicate-NAP extract have not all passed.

TL;DR: Buy Screaming Frog at £199/year to extract listing templates and sample the sitemap. Buy Sitebulb for hint reports a directory operator can read. Buy Lumar, OnCrawl, or Botify when logs and JS listing cards are the failure. Buy Semrush ($139/mo monthly SEO plan) or Ahrefs (Lite $129/mo) when you already need rank and crawl credits in one suite. Do not request indexing on thousands of listings until the template passes.

Why directories fail technical audits

Directories fail like marketplaces with worse uniqueness. User-generated listings clone NAP. City hubs paginate into five digits. "Claim this listing" URLs index. Two vendors occupy one address. A recrawl request then sits for days to weeks and still does not guarantee inclusion.

BEST_OF earn rate: 15.2% according to US Tech Automations (2026), on 12,514 live pages counted 2026-08-24. The online_directories vertical is not in the counted earn-rate table; Neutral vertical default: 10 according to the first-party mix-config (2026), not a vertical earn rate.

Paid spider licence: £199/year according to Screaming Frog (2026). Lite list: $129/mo according to Ahrefs (2026). SEO plan list: $139/mo according to Semrush (2026). Recrawl docs updated: 2025-12-10 according to Google Search Central (2025); that page still says crawling after a recrawl request can take a few days to a few weeks and is not a guarantee.

Wave-required siblings that are live: best online reputation management software for small business, local SEO for ecommerce stores, and local SEO for restaurants. Directory listing quality and local NAP problems overlap those pages even when the product is a directory, not a restaurant.

Key Takeaways

• Directory technical audits are template extracts plus pagination and duplicate-NAP, not a blog score.

• Dated public prices in this pass: Frog £199/year, Semrush SEO $139/mo monthly, Ahrefs Lite $129/mo (checked 2026-09-07).

• Recrawl is a hint. Google Search Central's recrawl doc (updated 2025-12-10) still refuses a clock.

• Ticket the listing template, not 25,000 listing 404s as 25,000 tasks.

• Zapier can ping sitemaps; it cannot own duplicate-listing policy unless you write that policy.

Glossary of crawl terms

• Listing template: Shared HTML for every business record. Extract NAP, category, and canonical from it.

• City hub: /tx/austin/ style pages. Thin hubs with 2 listings still index if you let them.

• Claim URL: /claim/123 should not be in the sitemap.

• Duplicate NAP: Two listing IDs, one phone. Models and Google both hate this.

• Pagination trap: Category page 400+ with no unique titles.

• Coverage state: Search Console's label; sample it, do not bulk-request 25,000 URLs.

• List mode: Frog crawl of a URL list (sitemap slice) instead of following every "related listing" link into a trap.

• Parameter sort: ?sort=rating should usually canonicalize or noindex.

Criteria

Feature matrix

Pricing

Vendor profiles

Screaming Frog

Best fit: List-mode a sitemap slice; extract NAP; find duplicate titles. Limitations: Full 25,000-URL JS crawl needs RAM and time; free cap is 500. Implementation: Custom extract phone and canonical; near-duplicate on listing body. Primary evidence: Screaming Frog. Who should not buy: Teams that refuse to run a laptop crawl.

Sitebulb

Best fit: Explaining duplicate clusters to an operator. Limitations: Contact vendor. Implementation: Hint reports on thin city hubs. Primary evidence: Sitebulb. Who should not buy: Data teams that only want CSV.

Lumar

Best fit: Staging vs production before a listing-template launch. Limitations: Contact vendor. Implementation: Fail CI when canonical extract empty. Primary evidence: Lumar. Who should not buy: 200-listing directories.

OnCrawl

Best fit: Logs that show Googlebot on page-400. Limitations: Contact vendor; needs logs. Implementation: Segment listing vs pagination vs claim URLs. Primary evidence: OnCrawl. Who should not buy: Directories without logs.

Botify

Best fit: Crawl-budget control on huge listing graphs. Limitations: Contact vendor. Implementation: Alarm when pagination out-crawls new listings. Primary evidence: Botify. Who should not buy: Pre-launch directories.

Semrush

Best fit: Rank tracking plus Site Audit when the suite is already the invoice. Limitations: $139/mo SEO plan is 5 websites and 500 keywords — not a log platform. Implementation: Track category hubs, not every listing. Primary evidence: Semrush pricing. Who should not buy: Teams that only need a desktop extract.

Ahrefs

Best fit: Crawl credits plus backlink context on city hubs. Limitations: Lite is 100,000 credits and 25,000 max pages per project on the pricing grid. Implementation: Cap the project to hubs + a listing sample. Primary evidence: Ahrefs pricing. Who should not buy: Teams that only need NAP extract.

Screaming Frog vs Sitebulb for this audit

Frog wins when the job is a list-mode extract of phone, canonical, and H1 at a published £199/year. Sitebulb wins when the operator will not open that CSV and needs a hint narrative. Neither one replaces Semrush or Ahrefs for rank time series, and neither one replaces OnCrawl for logs. If you can only buy one specialist crawler this quarter, buy Frog, run the 7 steps, and put Sitebulb on the next quote if the operator still cannot read the export.

Semrush vs Ahrefs on a directory is a suite decision, not an audit decision. Semrush SEO at $139/mo monthly ($117.33/mo annual) includes 5 websites and 500 keywords — enough to track hubs, not 25,000 listings. Ahrefs Lite at $129/mo includes 100,000 crawl credits and 750 tracked keywords. If you already pay for one suite, do not add the other "for the audit." Use the suite for rank, Frog for extract.

Hardware still matters. A 16 GB laptop will list-mode 2,000 URLs with JS on listing cards. A 4 GB office PC will die at 400 and you will think the directory is small. The free 500-URL cap is a product limit; RAM is a separate limit on the paid licence. Budget the machine or budget the cloud crawler. Do not budget neither.

Ops ownership is the rest of the audit. Duplicate NAP is a data problem. The crawler only finds it. Someone in listings ops has to merge records, and someone in SEO has to keep the loser URL as a 301, not a 404, if it had links. Put both names on the ticket. If ops will not merge, stop buying crawlers.

7-step directory audit

1. Inventory sitemaps: listings, hubs, cities. Remove claim and thank-you URLs.

2. List-mode crawl a sample (500–2,000 URLs) with NAP extract.

3. Cluster duplicate NAP and duplicate titles. Ticket the template, not each clone.

4. Write a pagination cutoff. Noindex or canonical beyond it.

5. If logs exist, join them. If not, do not buy OnCrawl "just in case."

6. Inspect 10 hub URLs in Search Console. Read coverage state. Do not bulk recrawl 25,000 listings.

7. Open tickets at template grain. Recrawl samples only after the template passes.

Worked example: a 25,000-listing directory with 400 categories and a $49/month listing upsell ships a new card component. Stripe invoice.paid fires for 180 new listings in 72 hours. A 3-hour paid Frog list-mode of 2,000 sitemap URLs plus Semrush Site Audit on the hub host shows 14% duplicate phones and 9 city hubs with identical 40-word intros. Search Console on a hub returns inspectionResult.indexStatusResult.coverageState = Crawled - currently not indexed. The ticket stays open until duplicate NAP has an owner, the 9 hubs gain unique facts, and a human ops lead signs that claim URLs are out of the sitemap. US Tech Automations would ingest the extract CSV, collapse duplicates by phone, retry failed inspections on 10 hubs only, and stop for human review before anyone uses the recrawl button as a publish step.

Directories also leak into reputation and local pack fights: a restaurant listed twice with two phones will not be saved by a crawler. That is why the live reputation and local-SEO guides linked above sit next to this audit, not instead of it.

When NOT to use US Tech Automations: If the directory is 200 listings and the founder can merge duplicates in the CMS, stay there. If Botify already alarms on pagination and ops honors it, do not add a parallel queue. If Make.com already pings the sitemap after invoice.paid and a human checks 10 hubs on Monday, staff that check.

Zapier, Make, and n8n can retry sitemap pings, log runs, and branch on errors. They will not invent a duplicate-NAP policy, idempotent ticket keys, or who is allowed to noindex /c/. A proposed ticketed design would key on template ID plus phone, collapse 400 clones into one ticket, and require ops + SEO dual approval before a robots change.

See dated prices and the ticket layer on pricing. Other workflows start on the company homepage.

Do not confuse a green Site Audit score with a clean directory. A 90 "health" with 3,000 duplicate phones is a failed audit. Put the duplicate count on the same Monday scoreboard as the health number. If finance only wants one number, give them duplicate-NAP count, not health.

City hubs need a unique fact the same way broker neighborhood pages do: a count of live listings, a last-updated date, and a category mix. If you cannot generate that fact from the database, do not publish the hub. Thin hubs are why recrawl requests sit for weeks and still miss.

Claim-this-listing URLs are a second class of waste. They are useful to sales and poisonous to the sitemap. Crawl a sample of /claim/ paths; if they are 200 and indexable, ticket the template for noindex plus sitemap exclusion the same day. Thank-you URLs after a claim are a third class. Three URL types, three rules, one Monday export. Do not debate them as "maybe content."

User-generated photos and hours go stale. The audit should extract lastmod or an equivalent field when the CMS has one. If lastmod is the crawl date for every listing, the sitemap is lying and Googlebot will learn to ignore it. Fix lastmod in the generator before you buy more crawl credits. Credits spent on a lying sitemap are not an SEO program. They are a bill. Ahrefs Lite at $129/mo will happily spend 100,000 credits recrawling that lie. Semrush Site Audit will happily score it. Frog list-mode on 500 URLs will show the lie cheaper. Start there, then spend credits on the hubs that already have unique facts and clean NAP. Everything else waits until the template ticket is closed by a named owner in listings ops and SEO on the same ticket, not in two separate Slack threads that nobody closes before the next scheduled weekly crawl run tonight.

Questions

Which tools belong in a directory technical SEO audit?

Frog or Sitebulb for extracts, Semrush or Ahrefs if you already pay for a suite, OnCrawl/Botify/Lumar when logs or JS exceed a laptop.

Is Screaming Frog or Sitebulb better for directories?

Frog if you want list-mode and custom extract at £199/year. Sitebulb if the operator needs hints. Many teams keep both.

How long after a recrawl request will listings index?

Google Search Central still says a few days to a few weeks, and not at all in some cases (doc updated 2025-12-10). Fix the template first.

Should every listing be in the sitemap?

Live, unique, indexable listings yes. Claim URLs, suspended listings, and pure duplicates no.

Can n8n replace Lumar?

n8n can ping sitemaps after payment events. It cannot be staging QA or a render crawl unless you build and maintain that.

When is Botify worth it?

When logs prove pagination is starving new listings and procurement can finish. Sample crawls come first.