Panda 4.0 and the Pages Nobody Reads

A new Panda, confirmed yesterday

Google confirmed yesterday that it has begun rolling out Panda 4.0, a new version of the update it uses to judge the quality of content across a site.

Owners who watch their visits daily may see a change this week, up or down, and wonder whether anything on their own site played a part.

Panda is not about links or technical errors. It asks whether a site's pages are worth reading, and on many small business sites the honest answer for some pages is no.

What Panda looks at

Panda first appeared in early 2011, aimed at sites with large numbers of thin, copied or low-value pages. It has been refreshed many times since, usually without announcement; this time Google named the version.

Google has said since Panda's first year that low-quality pages in one part of a site can affect how the whole site does in search. That is the part that matters to a small business.

A site with thirty good pages and three hundred empty ones does not look like a site with thirty good pages. It looks like a site that is mostly empty.

Pages that fit the description

• Tag pages created automatically for every tag, each listing one or two posts.
• Empty or nearly empty categories in a shop, showing no products or one.
• Product descriptions copied word for word from a supplier's catalogue and shared with every other dealer.
• Location pages that repeat the same text with a different town's name.
• Attachment pages that WordPress creates for each uploaded image, holding the image and nothing else.
• Old event and news pages with two lines and a date long past.

None of these was made to deceive anyone. Most were created by default settings or good intentions, and nobody has looked at them since.

Finding the pages nobody reads

Start with how many pages Google has indexed. The Index Status report in Webmaster Tools shows the count over time; compare it with the number of pages you believe the site has.

A large gap is informative. Far more indexed pages than expected usually means tags, archives, attachments or filtered product lists that nobody thought of as pages.

Then check which pages Analytics records as landing pages over the last twelve months. Pages that never received a visit from search, or never received a visit at all, are the first candidates.

Merge, improve or remove

Each weak page gets one of three answers. If it covers the same subject as a stronger page, merge it: move anything useful across and redirect the old address to the new one.

If it covers something customers genuinely need, improve it: write the content the page was always meant to hold, in your own words.

If it serves no reader, remove it, or keep it for visitors and tell search engines not to index it.

Automatically generated pages, such as tag and attachment pages, can usually be switched off or kept out of the index in the site's settings.

Copied product text

Shops that sell the same products as many others often use the supplier's description unchanged. Google then sees the same paragraph on dozens of sites and has no reason to prefer yours.

Rewriting every product at once is rarely practical. Start with the products you sell most often, and add what only you can say: fit, compatibility, what customers ask, and how you use it yourself.

Products that never sell and carry no description of your own may serve the site better as lines in a category list than as hundreds of separate thin pages.

A worked example

An outboard motor repair shop on a lake services engines from spring to fall and sells parts online, with about nine hundred product pages imported from a supplier's catalogue.

Each part carried the supplier's description, shared with every other dealer, and the category tree had dozens of empty branches. The shop's blog added a tag page for each of two hundred tags.

The shop's few genuinely useful pages, on winterising, spring tune-ups and common faults, were buried among the rest.

The developer kept tag and attachment pages out of the index, removed the empty categories and redirected them to their parents. The mechanic wrote new descriptions for the forty most ordered parts, naming the engines each one fits.

Parts that had not sold in two years were grouped into category lists, and the service pages were expanded and linked from the home page.

What not to do

Do not delete half the site in a weekend. Removing pages that do bring visitors, or breaking addresses that other sites link to, creates a new problem without solving the old one.

Do not reword copied text with a thesaurus or software that swaps words. The result reads badly to customers and is still recognisably the same text.

And do not pad thin pages with filler to reach some length. A short page that answers a question well is not thin; a long page that says nothing is.

Changes take time to show

Google has to recrawl the changed pages and reassess the site before improvements count, and Panda's assessments are refreshed from time to time rather than daily.

Make the changes for customers first. A merged page that answers a question fully, or a description that says which engine a part fits, is better for the person reading it today.

Keep a dated note of what was changed, so any later movement can be matched against it.

The counter-case

A small brochure site of fifteen pages, each written by the owner about a real service, has little to fear from Panda and little to change.

And a change in visits this week may have nothing to do with Panda: the season, a competitor's advertising or a change on the site itself can all move the numbers.

But almost every site that has run for a few years has gathered pages nobody reads. Reviewing them is worth doing on its own merits, whatever Panda decides.

Sorting the weak pages

1. Compare the indexed count with the pages you know of.
2. List pages with no visits in a year.
3. Switch off automatic pages nobody needs.
4. Merge overlapping pages and redirect the old addresses.
5. Rewrite descriptions for your best-selling products.
6. Improve the pages customers genuinely need.
7. Check again in Webmaster Tools after a month.

Step two turns a vague worry into a list. Seeing which pages have gone unread for a year makes most decisions obvious and keeps the effort on pages that matter.

Rewriting the titles of the pages you keep, so the business and the service show before the cut-off, is covered in page titles cut shorter in the new search results.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.