Duplicate page consolidation

What duplicate page consolidation is

Duplicate page consolidation takes a named set of pages that say much the same thing and turns them into one page, then redirects the old addresses to it and repoints the links that used to lead there. It is done to a specific set of pages you can list, usually between two and five of them, on one subject.

It is not the site-wide plan of what should merge, which looks at everything you have published and decides which clusters need collapsing. That is the survey. This is the work on one cluster. If you already know that your three cleaning pages are the problem, you do not need the survey first.

It is also not a page rewrite. A rewrite starts from one page that is wrong. This starts from several pages that are each partly right, and the job is deciding what survives from which.

When you need one

You need this when readers land on the wrong version. Someone arrives at the page written three years ago, sees an old price and a service you stopped offering, and never sees the good page two clicks away. Your own staff do it too: three people send three different links for the same enquiry.

The second sign is pages competing with each other for the same reader. Each has a few internal links pointing at it, each takes a share of the visits, and none is strong enough to be the page anyone remembers. Together they would be one substantial page. Apart they are three thin ones.

You do not need this if the pages only look similar and actually serve different buyers, which happens with industry pages and location pages more often than people expect. Merging those loses you something. Nor do you need it if the duplicates are worthless rather than overlapping, in which case deleting is faster and cheaper than merging, and that is what pruning is for.

What goes in it

• A side-by-side read. Every page in the set read against the others, with a mark against each paragraph: unique, duplicated, contradicted, or out of date. The contradictions are the interesting part, and there are usually several.
• The decision on which address survives. Normally the page with the most links pointing at it and the clearest address, not the newest or the longest. The reasoning is written down, because somebody will ask later.
• The merged page. Written as one piece rather than stapled together, keeping the best passage of each original and resolving every contradiction in your favour after checking the facts with you.
• The redirect list. Each retired address and the address it should point at, in a form your developer or your platform can act on directly.
• The internal links. Every place on your own site that points at a retired page, listed so it can be repointed rather than left to bounce through a redirect.

What you get

The merged page. One editable document, with a note in the margin of each section saying which original it came from, so you can satisfy yourself that nothing you valued was dropped.

The retirement list. Old address, new address, and a one-line reason for each. This is the document you hand to whoever controls the site, and it is written so they need no further explanation.

The contradiction note. A short list of the places where your own pages disagreed with each other on price, coverage, timing or what is included, and what each one now says. Businesses are usually more surprised by this than by anything else in the job.

A worked example

An illustration. A Calgary commercial cleaning company has three pages describing the same crew doing the same work: “Office cleaning”, “Commercial cleaning services” and “Janitorial services”. They were written in different years by three different people. Two mention a nightly service and one does not. One lists carpet extraction, which the company stopped offering when the machine was sold. One carries an old office phone number. Each has a handful of links pointing at it from elsewhere on the site, and enquiries arrive from all three.

The merged page keeps the office-cleaning address, because eleven internal links and the company's own email signature already point there. It opens with a plain statement of what the crew does and when, takes the nightly-service section from the second page because it was the clearest writing of the three, and drops carpet extraction with a single line naming the firm that work now goes to. The other two addresses redirect. Nine internal links are repointed and two deleted, because they sat in a paragraph that no longer made sense. The job takes eight working days, most of it spent settling which of three descriptions of the overnight schedule was correct.

How it runs

1. A half-hour call. Which pages are in the set, which one you think is best, and whether anybody has a reason to keep one of the others alive. No charge.
2. A fixed price in writing. Per cluster, agreed before anything starts, with the number of pages in the set named in it.
3. Access. The pages themselves, a list of where they are linked from, and someone who can answer questions of fact when the originals disagree.
4. The work. One to two weeks for a set of three to five pages. Most of that is the side-by-side read and the checking, not the writing.
5. Delivery. The merged page, the retirement list and the contradiction note, with one round of changes. Putting the redirects in place is your developer's job, or handled as a website update.

What it costs

Quoted as a fixed price in CAD after the half-hour call, because clusters vary enormously. Three short pages that mostly agree is a few days. Five long pages written by five people who each believed they were describing something different takes considerably longer, and the checking is where the time goes. The pricing page shows the four ways of working. The number is fixed in writing before anything starts.

What happens next

Three things generally follow. If this cluster turned out to be one of several, the site-wide survey is the sensible next purchase, and clearing years of accumulated pages in one pass is the version that includes the doing. If nothing on the site points at the merged page yet, a sweep for links that no longer go anywhere is the cheap follow-up. And if the contradiction note showed that your pages disagree because nobody owns them, that is a governance problem rather than a writing one. None of it is required. Merging one cluster is a complete piece of work, and one page where there were three is a finished result.