Skip to content
Get Ranked On TopIllustrative case studyCommerce SEO plan

A fictional business. The figures are a modelled scenario, not a client result.

Spec sheet GC-13 · Sheffield workshop

Grindle Cycles · Bike shop and online parts store, 6,000 SKUs · Sheffield

How a Sheffield bike shop trued 6,000 part pages and took its category rankings back

+64% organic revenue

13 months · UK & Ireland · Commerce SEO plan · Variant canonicals · Out-of-stock handling · Unique spec sheets · Product schema

Lateral runout, month 1 → month 13

Runout 2.5 mm → 0.2 mm

Client

Grindle Cycles

Catalogue

6,000 SKUs

Period

13 months

Status

Modelled, not a client

Plan

Commerce SEO

§01The fault

Six thousand parts, rolling out of true.

Every groupset size had its own near-identical part page and out-of-stock items returned 404s, so thin duplicates dragged the shop's category rankings down.

Grindle Cycles fixes bikes in a Sheffield workshop and ships 6,000 parts across the UK. Its platform gave every SKU a page of its own, so each cassette range, chain length and tyre width became a near-copy competing with its siblings, and anything that sold out simply vanished. Like a wheel with a dozen slack spokes, the shop still rolled, but the category pages that should have carried it rubbed on every turn.

One 12-speed cassette, nine URLs

Same photo, same copy, one number changed. Across the catalogue: 4,380 variant URLs, and 1,120 part pages answering 404 at any one time.

§02Inspection

What the stand measured.

  1. Finding 1 of 4

    Variants

    Measured

    4,380 URLs

    Tolerance

    0 indexed duplicates

    Every gear range, chain length, tyre width and crank length had a URL of its own: 4,380 pages that differed from a sibling by a single number. Google picked one at random to rank, and the category pages linked to all of them.

  2. Finding 2 of 4

    Stock

    Measured

    1,120 × 404

    Tolerance

    0 for parts coming back

    The platform unpublished an item the moment stock hit zero. At any one time 1,120 part pages answered 404, 310 of them with links from forums and build guides, and they came back days later with no rankings.

  3. Finding 3 of 4

    Specs

    Measured

    38 words

    Tolerance

    ≥ 250 words of its own

    The median product page carried 38 words of its own. The rest was the manufacturer's description, word for word the same as on dozens of other UK shops.

  4. Finding 4 of 4

    Categories

    Measured

    #19

    Tolerance

    Top 10

    Category pages for searches like “12 speed cassette” averaged position 19: their internal links and authority were spread across thousands of variants and dead ends.

Where Googlebot's requests went

Before · log-file sample
  • Variant duplicates 46%
  • 404s 18%
  • Product pages 22%
  • Category pages 14%
Month 13 · log-file sample
  • Variant duplicates 6%
  • 404s 1%
  • Product pages 61%
  • Category pages 32%

§03The build

Four parts on the bill of materials.

Built on our Commerce SEO plan: the ecommerce engagement for catalogues where variants, stock and product data decide what ranks.

  1. 01

    Variant canonicals

    Size and speed variants of the same part were folded into 1,620 parent pages with a variant selector. Variant URLs kept working, carried a canonical to the parent and left the XML sitemap; the few with real search demand of their own were kept as pages.

    4,380 variant URLs

  2. 02

    Out-of-stock handling

    Zero stock no longer unpublishes a page. Parts coming back stay live with a restock date, OutOfStock availability and in-stock alternatives that fit the same bike; discontinued parts get a 301 to their direct successor.

    1,120 part pages

  3. 03

    Unique spec sheets

    Manufacturer boilerplate replaced by a spec table built from the shop's own attribute data: weight, speeds, freehub and bottom-bracket standards, chain line, and a fitting note written by the workshop mechanics.

    1,480 spec sheets

  4. 04

    Product schema

    Product and Offer markup on every parent page, one offer per variant with its own price, GTIN and availability, so results show the right price and stock for the size searched. Category pages link to parents only.

    1,620 parent pages

§04On the truing stand

Each spoke is a product URL. Tighten them and the wobble goes.

Thirty-two spokes stand in for the catalogue: twelve near-duplicate variants, eight out-of-stock pages, six boilerplate spec pages and six that were already fine. Turn a spoke key to fix a whole kind, or pick one spoke to see what happened to that URL.

Pads rubbing

Runout (the metaphor's reading)Out of true

2.5 mm

Runout trace · one revolution±0.5 mm tolerance
-2-10+1+20°90°180°270°360°

Spoke keys

Each key tightens every spoke with that fault.

Variant duplicates indexed
4,380from 4,380
Out-of-stock 404s
1,120from 1,120
Boilerplate-only pages
1,480from 1,480
Organic revenue a month (modelled)
£38,500from £38,500

Spokes, one product URL each: choose one to inspect

Arrow keys move between spokes.

Spoke 1 of 32

/cassettes/12s-11-28

Near-duplicate variant

Before

200 · duplicate

A near-copy of its siblings: same photo, same text, one size changed. It competed with them, and with /cassettes/12s-road, for the same searches.

After

200 · canonical → /cassettes/12s-road

Still there for the size selector, but canonical to the parent and out of the sitemap. Its signals now add up on /cassettes/12s-road.

§05Month by month

13 months, link by link.

  1. Months 1–2

    Inspection

    Full crawl of 6,000 SKU URLs, twelve weeks of server logs, stock-status history and revenue per URL; every variant mapped to a parent.

  2. Months 3–5

    Canonicals

    Parent pages with variant selectors released category by category, cassettes and chains first; canonicals live, sitemaps cut to parents.

  3. Months 6–7

    Stock handling

    The zero-stock rule changed in the platform: restock dates and alternatives on live pages, 260 discontinued parts redirected to successors.

  4. Months 8–10

    Spec sheets

    1,480 spec sheets generated from attribute data and checked by the workshop, best-sellers first; Product and Offer markup on every parent.

  5. Months 11–13

    Category pages

    Category copy, buying guides and internal links pointed at parents only, so the consolidated authority turned into category rankings.

§06Results

13 months on, the wheel runs true.

Organic revenue

+64%

£38,500 → £63,100 a month

Near-duplicate variant pages indexed

4,380 → 0

Canonical to 1,620 parent pages

Out-of-stock 404s

1,120 → 0

860 kept live with a restock date, 260 redirected to a successor

Category keywords in the top 10

7 → 26

of 40 tracked category keywords

Modelled scenario for a fictional business — not a client result. Rankings and traffic are never guaranteed; see our Terms.

Organic revenue a monthGrindle CyclesNo changes (modelled)
£30K£40K£50K£60K£70KM1M3M5M7M9M11M13£63,100£35,800
  • Inspection
  • Canonicals
  • Stock handling
  • Spec sheets
  • Category pages

Category searches, average position

SearchMonth 1Month 13
12 speed cassette#23#6
11 speed chain#18#4
gravel tyres 700x45#27#8
bsa bottom bracket#31#7
resin disc brake pads#16#3
mineral oil bleed kit#22#5
road crankset 172.5#35#9
bike shop sheffield#4#1

Modelled monthly averages for UK searches. 26 of 40 tracked category keywords now sit in the top 10.

True your catalogue

Variants, stock and spec pages pulling your categories out of line? Start with the Commerce SEO plan, or tell us what your shop sells.

← All case studies