Should You Combine Two Websites After an Acquisition or Run Both?

What Actually Moves Between Two Sites

A permanent redirect is an instruction that one specific resource now lives at one specific new address. That is the whole mechanism, and its narrowness is the reason most consolidations disappoint. Nothing about the instruction operates at the level of a website. It operates at the level of a URL, which means the value attached to a page moves only if somebody wrote a rule for that page.

The distinction between permanent and temporary is not stylistic. For a permanent redirect the indexing pipeline treats the redirect as a signal that the target should be canonical. For a temporary one the crawler still follows the hop, but the pipeline does not use it as a canonical signal at all. So a migration served with 302 responses will look correct in a browser while transferring nothing, which is the single most expensive configuration error in this whole exercise.

The underlying standard is older than any search guidance and says the same thing. Section 15.4.2 of the HTTP specification defines 301 as meaning the target resource has been assigned a new permanent URI, and instructs clients to update their cached references. Section 15.4.9 defines 308 identically, with the added guarantee that the request method is preserved. Section 15.4.3 defines 302 as temporary, which is precisely why it carries no consolidation weight.

Two of the three end states are supported by the platform's own migration tooling. The third, the one that consolidates most efficiently, is not. That single asymmetry decides more consolidations than any argument about authority, and it is set out in detail further down.

Redirects are not the only instrument, and the instruments are not equal. Google describes redirects plus rel=canonical annotations as strong signals that a given URL should become canonical, a sitemap entry as a weak one, then notes that the methods stack to become more effective in combination. It also ranks the implementation methods by how likely it is to interpret them correctly, with a server-side response at the top and a JavaScript redirect near the bottom.

Read that table as a budget rather than a menu, then apply it one URL at a time. The instruction is per-resource, so a consolidation is not a switch thrown once at the site level. It is a decision taken separately for every address the acquired estate holds, and each of those decisions has exactly three available outcomes.

Read that table as a budget rather than a menu. A consolidation carried entirely by rel=canonical is a request; a consolidation carried by server-side permanent redirects is an instruction. Anything below the first two rows should be treated as a temporary measure with a date attached, because the further down the list the implementation sits, the more of the decision is being delegated to inference.

Which Estate Receives

The two websites are not symmetrical, and the vocabulary matters here. One of them is the recipient estate, meaning the address that survives and accumulates the combined signal. The other is the donor. Which is which is a question of measured position, not of which logo the board prefers, and getting that backwards is how a company spends a year redirecting a strong property into a weak one.

The temptation is to assume the parent brand wins by default. Often it does. But an acquired specialist with fifteen years of category-specific coverage can hold a stronger position on the queries that produce revenue than the group brand that bought it, and the group brand will still have better name recognition. Recognition is not retrievability, and only one of the two is being consolidated.

There is also a wider lesson in how rarely a declared preference turns out to be an earned one. Across the crawled web, roughly two thirds of pages carry a canonical tag, which is to say they state which version should win. Only a small fraction are actually the version that anything else points at.

The 7% figure is not a failure rate; most pages are self-canonical and that is correct. What it illustrates is the distance between an assertion and an outcome, which is exactly what a consolidation tests. Google is explicit that if no canonical is specified it will identify which version is objectively the best to show, which is a polite way of saying it reserves the right to disagree with you. A merge built on an assertion the evidence does not support will simply be overruled.

So the recipient is chosen from data before anything is written. Which estate holds position on the queries that produce qualified demand, which holds the references from other sites, which has the longer unbroken measurement history. The answer is sometimes uncomfortable, and it is much cheaper to be uncomfortable before the redirect map exists than after. A structured review of both properties is the same exercise described in deciding whether a website needs replacing, applied to two estates at once.

What the Purchase Agreement Decides

Before any of this is a technical question it is a documentary one, and the document is the purchase agreement. This is the step integration plans skip, because the people who read the agreement and the people who write the redirects rarely meet. The agreement decides what you are permitted to do with the acquired estate, and it decides it in language that is easy to verify.

A representative example is on public file. The May 2025 asset purchase agreement by which Beyond, Inc. acquired the Kirkland's brand assets defines a distinct category of acquired property: all domain names comprised of or containing the element Kirkland, or any abbreviation, translation, or derivation of it, including those set out on a numbered schedule. Domains are not swept up with the trademarks. They are enumerated, by name element, on a schedule of their own.

The agreement goes further in a way that should interest anyone who has ever tried to redirect a domain they did not control. Its intellectual property schedule requires that for each item, specifically including each internet asset, the parties list the registrant, the registrar, plus the account holder. A separate schedule lists every contract under which the seller licensed those same marks or domains to somebody else.

Every row marked yes in that table is a way for a consolidation to fail after it has been approved. A domain nobody scheduled was never acquired, so it cannot be redirected. A domain whose account holder is an agency rather than the company requires that agency's cooperation, on their timetable. A domain carrying a live third-party licence may not be lawful to retire at all, whatever the integration plan says.

The practical instruction is short. Get the schedules, reconcile them against the domains your team believes it inherited, and treat any discrepancy as a blocking item rather than a detail. This is the same class of check an acquirer runs in the other direction, which is set out in how a search position is credited in a sale.

The DSF Estate Consolidation Matrix

The question in the title presents two options. There are four, two of which involve keeping both sites for a stated reason over a stated period. The DSF Estate Consolidation Matrix produces the verdict from two questions: whether buyers reach the two brands through the same searches with the same intent, and whether the deal actually delivered control of the acquired estate.

Read the plane before the table. The two arrows say the only thing a list of four rows cannot: the vertical axis is a lever you can pull, because registrar access, Search Console ownership and the analytics history can all be obtained after the fact. The horizontal axis is not a lever at all. Whether buyers reach both brands through the same searches is a fact about buyers, so no amount of integration work moves an estate sideways.

Consolidate is the verdict when demand overlaps and control is complete. Name the recipient on evidence, map every donor URL to a real counterpart, serve permanent redirects, retire the donor, then hold the redirects past the signal window. Close the Gaps is the verdict when the demand overlaps but the control does not yet exist, and it is the most common correct answer in the first year after a deal. Run both deliberately, close the registrar, ownership, analytics, and source gaps, then consolidate at the next release.

Run Both, One Entity applies when the demand is genuinely distinct. Keep two addresses, because the searches are different, but declare a single organization graph across them so the entity layer is unified even though the URL layer is not. Ring-Fence applies when the demand is distinct and control is partial. Run both, keep the acquired estate independently operable, and accept that you are buying optionality, including the option to sell it again.

The value of separating these is that it stops the conversation being a preference argument. Two factual questions produce the verdict, both of them answerable in a week, and one of them answerable only by reading the deal documents rather than a dashboard.

Why a Folder Forfeits the Migration Tool

The most consequential constraint in this whole subject is buried in a support article, and almost nobody writing about post-acquisition consolidation mentions it. Google's Change of Address tool forwards signals from an old site to a new one for 180 days. It requires that you own both properties in a single Search Console account. And it operates only on properties at the domain level.

That last clause is the trap. The tool explicitly does not support moving pages within the same domain to different paths, and the documentation names the excluded case directly: a destination such as a folder beneath your existing site is not supported. So the pattern that consolidates authority most efficiently, folding the acquired brand into a subdirectory of the surviving domain, is exactly the pattern the migration tool cannot assist with.

Nothing about that makes a subdirectory the wrong answer. It shifts the entire burden onto the redirect map, which now has to be complete rather than merely good, because there is no supplementary signal underneath it doing quiet corrective work for six months. A subdirectory consolidation with a partial map has no safety net.

There is a discrepancy between two pieces of the same publisher's guidance, and it is worth stating plainly rather than smoothing over. The Change of Address documentation says to maintain redirects for at least 180 days, longer if traffic persists. The site move documentation says to keep them as long as possible and generally at least a year, so that all signals transfer. Take the longer figure as the operating rule, budget the hosting for it, and treat the 180 day window as the period in which active help is available rather than the period in which the work is finished.

That instruction is the death of the shortcut every under-resourced integration reaches for. Pointing the acquired domain's entire URL set at the new homepage is fast, it is a single rule, it looks tidy in a status report, and it discards the position of every page it flattens. The permitted version of the same maneuver is narrow: where content genuinely was consolidated from several pages onto one new page, the old URLs may point at that consolidated page. Consolidated content, not convenient content.

When Running Both Is Correct

Keeping two sites has a reputation as the indecisive option. It is not. The platform's own documentation says so in the context of regional targeting, where it lists easy separation of sites as a genuine advantage of separate domains or subdomains, then harder separation as a real cost of subdirectories. Separation has value. The question is whether you are buying anything with it.

Operators run multi-brand estates at meaningful scale. Bed Bath & Beyond, Inc. reported first quarter 2026 net revenue of $248 million, up 6.9% year over year, and described it as the first significant revenue growth in nineteen quarters. The same release describes the company as an ecommerce-focused retailer with an affinity model that owns or holds interests in various retail brands, naming Bed Bath & Beyond, Overstock, buybuy BABY, and more recently the Kirkland's brands. Four distinct consumer propositions, four distinct search intents, held deliberately apart.

That is the case for running both stated properly: the demand really is different, so a single address would serve every one of those audiences slightly worse. It is not the case for running both because nobody made a decision, which looks identical from the outside for about three quarters and then starts to cost money.

The cost of the keep-both choice is mostly arithmetic. Two full URL sets is how a medium estate becomes a large one, and crawl budget becomes a live concern once a site passes a million pages changing weekly, or ten thousand pages changing daily. Google is direct about the consequence: time spent crawling URLs it should not be crawling means it might not explore the rest of the site, or might not increase the budget. Two half-maintained estates can therefore be worse than one, which is covered in more depth in optimizing crawl budget at scale.

So run both when the searches differ, and write down which searches those are. If the list is short, or if it is really a list of internal preferences rather than queries, the honest verdict is Close the Gaps rather than Run Both, and the difference between those two is a date.

How the Knowledge Graph Merges

Consolidation does not stop at the web server. The structured record of who your company is has to merge too, and it merges on remarkably similar terms. Wikidata's merge procedure is worth reading by anyone planning a brand consolidation, because it describes the same asymmetry with none of the ambiguity.

A merge happens when two records exist for the same topic. The surviving record is usually the one used more often, or the older of the two. Labels, descriptions, aliases, sitelinks, plus statements are consolidated into the survivor, and the obsolete record becomes a redirect to it. Identifiers are persistent, so a merged record must never be reused for something else. And there is a warning that will be familiar to anyone who has watched a website migration go wrong: do not move the claims manually, because if you do, the redirect is never created.

Copying the acquired brand's content onto the surviving domain without redirecting the old URLs is the same mistake as moving the claims by hand. Both leave two records describing one thing, with no instruction about which is current, and both invite a resolution you did not choose. That is how a merged company ends up described by an answer engine in the past tense, or described twice, with different facts.

What can be declared explicitly is narrower than most people assume. Schema.org gives an organization parentOrganization plus subOrganization for hierarchy, brand for the brands it maintains, sameAs for unambiguous external identity, alternateName, then legalName. It gives no property for a merger, a predecessor, or a former name. There is no tag that says this company used to be that one, so the continuity story has to be carried by the graph you assert across both estates and by the redirects that back it up, which is the practical reason the Run Both, One Entity verdict insists on a declared parent.

What the Deal Calendar Allows

Every deadline in this exercise is set by two clocks that are not aligned. One is the deal calendar, which is measured in years. The other is the signal window, which is measured in months. Integration plans routinely assume the first governs the second.

Bain reports that deals above $10 billion take roughly seven months from announcement to close, then another 24 to 36 months to realise the bulk of run-rate cost synergies, and that large deals frequently take 36 months or more from announcement to full integration. Against that, the active forwarding window is six months and the recommended redirect floor is a year. The website decision cannot wait for the integration programme to reach it, because by then the window that would have helped has closed.

The sequencing that follows from those two clocks is unglamorous. Lower the DNS time-to-live to a few hours at least a week before any cutover. Expect a temporary drop in crawl rate immediately after launch, followed by a steady recovery over the following days, and do not treat that dip as a failure. Monitor traffic on both origins, then retire the old hosting only once its traffic has reached zero. For a site moving without URL changes that is the entire procedure, and it is also the safest rehearsal for one that is.

The scale of the population making this decision is easy to underestimate, because the coverage concentrates on the largest transactions. Global deal value reached $4.9 trillion in 2025 and $2.4 trillion in the first five months of 2026, up 41% year over year, on pace to top $5.3 trillion.

The distribution matters more than the total. Deloitte's January 2026 survey of 1,500 dealmakers found more than 80% expecting to transact a greater volume of deals, and recorded that one third of all United States deal value in 2025 came from just twenty very large transactions. If twenty deals account for a third of the value, the overwhelming majority of transactions, and therefore the overwhelming majority of two-website decisions, are happening well below the headlines.

None of this is glamorous work, and almost none of it is marketing work. It is schedules from a purchase agreement, registrar records, Search Console ownership, a URL-level map that somebody actually finished, plus a hosting commitment long enough to outlast the signal window. The reason it keeps going wrong is not that it is difficult. It is that the decision gets made by whoever is available in month three, and the constraints that would have changed the answer live in documents they were never sent.

The decision is also reversible only at full cost. A consolidation can be undone, but the second migration pays the same price as the first and starts from a weaker position, so the cheapest version of this project is the one where the recipient is chosen once, on evidence. What that scoping looks like in practice is set out in what belongs in a website redesign scope.

For an evidence-led read on which of two estates should survive, and what each one would forfeit in a merge, a Website Health Audit is where that assessment gets built.

FAQ — Combining Two Websites

Next Steps — Combining Two Websites

Choosing a recipient estate on measured evidence, then scoping the map that carries the other one across, starts with Search Engine Optimization (SEO).