Most law firm websites that have been around for fifteen years have been rebuilt twice. Once around 2014, when the old table layout finally became indefensible, and once again more recently when the firm moved to a new platform or a new marketing vendor. Each rebuild came with a redirect map, usually built in a hurry during launch week, and each map was written against whatever the site looked like at that moment. Nobody went back and reconciled the second map against the first one.
The result is redirect debt. Your 2011 practice area URL points to your 2014 URL, which points to your 2021 URL, which points to a page that no longer exists because someone consolidated three car accident pages into one last year and did not tell the developer. A visitor clicking a link from an old bar association directory takes four hops and lands on a 404. Search engines do the same thing, get less patient about it, and quietly stop passing along whatever authority that old URL had accumulated.
This is unglamorous work. It does not produce a slide anyone wants to present. It is also one of the few technical fixes where the upside is concrete and the downside of ignoring it compounds every time the site changes again.
What a Redirect Chain Actually Costs
A single 301 redirect is fine. It is a supported, permanent instruction, and search engines handle it well. The problem starts at hop two.
Three costs stack up. The first is speed. Every hop is a full round trip to your server before a single byte of the destination page renders. On a decent host that is 80 to 200 milliseconds per hop, and on mobile over a weak connection it can be 300 or more. A four hop chain can add over a second of pure waiting before the page even starts loading, which lands directly on Time to First Byte and drags Largest Contentful Paint with it. Google’s own guidance on Time to First Byte names redirects as one of the primary contributors to slow server response, and recommends avoiding them wherever possible.
The second cost is crawl behavior. Googlebot will follow a chain, but it does not follow it forever. Roughly five hops in a single crawl attempt is the practical ceiling before the crawler gives up on that pass and queues the URL for a later attempt. Long chains also burn crawl budget on URLs that produce nothing. On a 400 page firm site that rarely matters. On a firm with a decade of blog archives, thousands of tag and date pages, and two generations of dead attorney bios, it very much does.
The third cost is the one partners care about. Links that other sites earned for you point at the start of the chain. If that chain ends in a 404, the value of the link is gone. Not diminished, gone. A directory listing from a state bar, a citation from a local news story about a verdict, a link from a law school alumni page, all of it evaporates at the moment the chain breaks.
Inventory Before You Touch Anything
The single most common way this project goes wrong is a well meaning developer opening the redirect plugin and starting to clean up. Do not edit anything until you have a complete list of URLs that have ever existed on the domain. There are five sources and you want all of them.
- The current site crawl. Run Screaming Frog or Sitebulb against the live site in spider mode. This gives you the set of URLs that currently exist and are linked from somewhere.
- Search Console. Export the Pages report, all of it, not just the indexed tab. The Not Indexed section is where redirect debt lives, specifically the Page with redirect, Not found 404, and Alternate page buckets. Also export the last sixteen months of Search Results data by page.
- Your existing redirect rules. Whatever they live in. A Redirection plugin export, .htaccess, an nginx config, Cloudflare bulk redirects, or all four at once. Firms that have changed hosts often have redirects in two places that contradict each other.
- Server access logs. Thirty to ninety days of raw logs will show you what real crawlers and real humans are still requesting. This surfaces URLs that appear in no crawl and no sitemap but still receive traffic from an old PDF, a printed ad, or an email signature from 2017.
- Backlink data. Pull the Best by Links report from Ahrefs, plus the Links report from Search Console. You want every URL on your domain that any external site points at, dead or alive.
Deduplicate all of that into one column in a spreadsheet. For a firm with two rebuilds behind it, expect somewhere between 800 and 6,000 unique historical URLs. Wide variance, and it depends almost entirely on how much blogging happened and whether the old CMS generated date archives and tag pages.
Crawl the List and Read the Result
Take that deduplicated column and run it through Screaming Frog in list mode with the setting to always follow redirects turned on. Then export the Redirect Chains report. That report is the entire diagnosis in one file. Each row shows the requested URL, every hop it takes, the status code at each hop, and the final destination and its status.
Sort by number of hops descending and read the top of the list. Then sort by final status code and separate the rows into four piles.
- Clean single hops. One 301 landing on a 200. Leave them alone. On most old firm sites this is only half to two thirds of the rows.
- Chains ending at a 200. Two or more hops, live destination. These work but they leak speed and crawl efficiency. Flatten them.
- Chains ending at a 404 or 410. The actual bleeding. Every one of these needs a decision.
- Loops and mixed status types. A 302 in the middle of a 301 chain, or a URL that redirects back to something earlier in its own chain. Loops are usually caused by an old rule fighting a new one, and they will hang a browser.
One more filter worth adding. Flag any redirect that ends on the homepage. Bulk homepage redirects are the classic launch week shortcut, and they are treated as soft 404s when the destination has nothing to do with the original page. A 2013 page about wrongful death damages redirecting to your homepage is not a redirect, it is a dead end with extra steps.
Flatten Everything to a Single Hop
The rule is simple and there is no exception to it. Every historical URL redirects directly to its final live destination in one 301. Not to the intermediate URL from the last rebuild. To the page that exists today.
Mechanically this means rewriting the source of each rule so the target is the current URL, then deleting the intermediate rules that are no longer reachable from anywhere. Two cautions. First, do not delete intermediate rules until you have confirmed nothing external points at them, because an intermediate URL from the 2014 site may itself have picked up links. If it has, it stays in the list as its own single hop entry. Second, keep the rule set in one system. If you have redirects in both a WordPress plugin and .htaccess, consolidate to whichever layer sits closest to the server, because it resolves faster and it is far easier to audit a single file than two systems whose order of operations nobody can explain.
Use 301, not 302, for anything permanent. Temporary redirects on permanent moves are common on firm sites that were migrated by a developer who did not know the difference, and they keep the old URL in the index indefinitely.
Also check protocol and host consistency while you are in there. A surprising number of old firm sites still resolve http to https and non www to www as two separate hops, which means every single visitor arriving from an old link pays a double redirect before your real rules even run. Collapse that into one rule at the server or CDN level. This is the same discipline that governs a full platform move, covered in more depth in our guide to law firm website migration without losing rankings.
Find the Old URLs Still Earning Links
This is where the money is, and it is the step most cleanup projects skip.
Take your backlink export and join it against the crawl results on URL. What you are looking for is any URL with referring domains pointing at it whose chain ends in a 404, a loop, or an irrelevant homepage redirect. Sort that joined list by referring domains descending and work down from the top.
On a fifteen year old firm site, the pattern is predictable. The pages with the most external links are almost never the current practice area pages. They are old blog posts about a change in state law, a scholarship page the firm ran for three years and then quietly killed, an attorney bio for someone who left in 2019, a resource page or state statute explainer, and occasionally a press release about a verdict. It is common to find twenty to sixty distinct URLs with real referring domains sitting behind broken chains. Reconnecting those is usually the highest value hour in the entire project.
For each one, pick the closest live equivalent by topic and intent. A dead scholarship page goes to the current community involvement or firm news page. A departed attorney bio goes to the practice area they handled, not to the generic attorneys index and definitely not to the homepage. An old statute explainer goes to whichever current page covers that topic. If nothing on the current site is a genuine match, the honest answer is often to rebuild the page rather than redirect it, particularly if it still has a dozen referring domains and receives measurable traffic. Recreating a well linked resource page is cheaper than earning those links again.
Decide What to Stop Redirecting Entirely
Not every dead URL deserves a rule. Redirect maps that never delete anything become their own performance problem, and a firm carrying 4,000 rules in a plugin is paying a database lookup on every request for no reason.
Retire the rule and let the URL return 410 Gone when all of these are true. No referring domains. No impressions or clicks in the last sixteen months of Search Console data. No requests in ninety days of server logs beyond bots. No current page that is a genuine topical match. That describes tag archives, date archives, paginated blog listings, old event pages, expired job postings, and the enormous tail of thin posts that the 2014 vendor published to hit a monthly quota.
A 410 tells search engines the removal is intentional, which gets the URL dropped from the index faster than a 404. Neither one harms the rest of your site. Dead URLs that were always worthless do not become valuable because you pointed them somewhere.
The tradeoff to be honest about. Deleting a rule is irreversible in practice, because once the URL is gone from the index you will not notice if you were wrong. So keep the full pre cleanup export permanently, with a date on it, and set the bar for deletion high. When in doubt, keep the single hop redirect. It costs almost nothing.
What Changes and What Does Not
Set expectations before you start. Flattening chains does not create rankings. It removes friction and reconnects value that was already earned and is currently being wasted.
What you should reasonably see, over four to twelve weeks as the site is recrawled, is a shrinking Page with redirect count in Search Console, fewer crawl requests wasted on non pages in the Crawl Stats report, faster server response on pages that were being reached through chains, and gradual recovery of visibility on pages sitting downstream of previously broken links. If the firm has a genuinely large number of well linked dead URLs, the recovery can be noticeable. If it has three, it will not be, and anyone telling you otherwise is selling something.
Redirect debt is also rarely the only thing wrong on a site this old. It sits alongside duplicate practice area pages, missing canonical tags, orphaned content and template level issues, which is why we treat it as one line item inside a broader technical SEO program for law firm websites rather than as a standalone fix. If your traffic has dropped and you are trying to work out whether redirects are the cause or a symptom, work the problem in order using our diagnostic order for law firm website traffic drops before you start editing rules.
Keeping the Debt From Coming Back
The reason firms end up here twice is that redirect maps are treated as launch artifacts instead of living records. Four habits prevent the third round.
- Keep one master redirect file in one system, exported to a shared spreadsheet after every change, with a date and the name of whoever made it.
- Any time a page is deleted, merged, or has its slug changed, the redirect is written in the same work order. Not later, not in a cleanup ticket.
- When you write a new rule, check whether the new destination is already a redirect target somewhere. If it is, update the old rules to point at the new destination at the same time. This one habit is what prevents chains entirely.
- Run the list mode crawl once a quarter against the master file. It takes twenty minutes and it catches drift before it becomes a chain.
None of this requires a new platform or a rebuild. It requires someone owning the file.
Pay This Debt Once
Redirect cleanup on an old firm site is a bounded project. Inventory and crawl in a day or two, decisions and rule rewrites in another two or three, verification the week after. What you get is a site where every link anyone has ever earned for the firm still lands somewhere real, and a rule set the next developer can read. Rubiks handles this as part of technical work on established law firm sites, including the backlink reconciliation step that most cleanups skip. Talk to Rubiks and we will crawl your historical URL set and show you exactly how many hops and how many broken links you are carrying before you commit to anything.