Google Shopping for technical products: why the feed decides everything

Two business types, two sets of problems

Agricultural eCommerce splits into two distinct models, and understanding which one you run determines where to look first.

Parts and machinery dealers carry large, complex catalogues. Tractor spares, machinery components, plant equipment parts. The buyer has specific intent. They know what machine they have and what has failed. Conversion is won or lost on whether your product page gives them three things: compatibility confirmation, specification detail, and a clear stock position. The problems here are structural.

Agricultural supplies businesses sell simpler, consumable ranges. Electric fencing, livestock ear tags, animal health products, farm sundries. The buyer is often habitual. They have bought from you before and will buy again, if you make it easy enough. The problems here are commercial: repeat purchase, Google Shopping efficiency, product feed quality, and retention.

The path to converting them, retaining them, and measuring it accurately is different enough to warrant distinct thinking. But both models share one foundational failure. The data they are making decisions on is unreliable.

Where parts and machinery businesses lose sales

Product discovery fails before the product page does

Most parts catalogues are organised around how the business stocks its inventory. Buyers do not search that way. They search by machine, by model year, by application. If your category structure does not reflect that, buyers cannot find the product even when you carry it. The sale is lost before it had a chance.

Site search compounds this. A buyer typing ‘3CX hydraulic filter’ into your search bar gets zero results when the product is catalogued as ‘Hydraulic Filter, JCB 3CX Series.’ Same product. Different language. The search tool cannot connect them.

The product page does not close the sale

Technical buyers arrive with specific intent. They know the machine. They need to know the part fits, what the specification is, and whether you have it now. Most parts product pages fail on all three. Generic manufacturer copy. One image. A stock status that says ‘usually dispatched in 3 to 5 days.’

That buyer closes the tab and tries the next result. It was not a pricing problem or a trust problem. It was a product page problem.

Paid channels spend on the wrong things

Performance Max campaigns run on autopilot and spend where the algorithm wants to spend, not where the commercial return is. Without clean conversion data feeding it, the algorithm optimises for the wrong signals. Performance Max has no keywords and no match types. It decides which searches to show for based on your feed and your assets, which means a thin feed produces broad, low intent coverage, and searches for your own brand name get picked up by a campaign you are paying prospecting rates for. Google has added campaign level negative keywords and brand exclusions to Performance Max, so both of these can now be controlled. Most accounts have never set them.

The campaign looks active. The qualified enquiry count sits still.

Where agricultural supplies businesses lose sales

Google Shopping campaigns: structured for volume, not margin

Supplies businesses running Google Shopping often have two problems at once. The product feed is not strong enough to compete on high intent terms, so the campaign defaults to broader, lower intent traffic. And the campaign structure gives no visibility into where the spend is actually going.

Elogic Commerce’s 2026 benchmark analysis puts B2B session to purchase conversion in a range of roughly 1.8% to 3.0%, and notes that many published B2B figures measure enquiries rather than completed orders. A supplies business selling to habitual buyers who know the product and are reordering should expect to sit at the upper end of that range. A campaign generating traffic that converts well below it is not a campaign problem. It is a feed and targeting problem.

Repeat buyers are not coming back through the store

Agricultural supplies customers are habitual. A farmer who bought ear tags in March will buy them again. A smallholder who fenced a paddock in spring needs more insulators in autumn. These are not win back situations. They are retention failures: a buyer who is ready to purchase from you again, and finds a competitor faster.

Most supplies stores have no post purchase communication, no reorder prompt, no reason built into the experience for the buyer to return directly rather than starting from a Google search.

Checkout friction on commodity lines

A buyer adds electric fence tape, an energiser, and a pack of posts to the basket. They reach the checkout. Payment is by credit card or PayPal only. No trade account option. VAT is not clearly shown. They call the merchant instead.

The sale happens off platform, credited to no channel, invisible to any reporting.

The buyers who do complete the order can go missing too. A checkout that sends the buyer off to PayPal or Opayo to pay and then brings them back will, on many setups, credit the order to the payment gateway rather than the campaign that brought the buyer in. The order is counted. The channel that earned it is not.

The problem both types share: the data is unreliable

GA4 is installed on most agricultural eCommerce stores in the UK. Very few have it configured to tell them anything useful.

Standard GA4 implementation does not automatically track eCommerce purchases. Purchase events fire on page loads rather than on confirmed transactions. Purchase events are missing entirely. A checkout step fires twice on mobile and doubles the count. Spam sessions contaminate the traffic data. The numbers look authoritative. They do not reflect reality.

The farming supplies business spending £1,500 a month on Google Shopping and seeing 45 purchases in GA4 may have 12 actual orders. The other 33 are tracking artefacts. If those GA4 key events are what has been imported into Google Ads as the conversion action, every bid decision Google’s algorithm is making is based on numbers that do not exist. And every strategic decision the business owner is making sits on the same fiction.

Fix the data before touching the channels. Wrong data poisons every decision that follows it. The second piece in this series covers exactly what breaks in GA4 and how to test it.

Three things to check this week

Regardless of which type of business you run, these three checks take less than an hour and will tell you whether your store has a data problem.

1. Check GA4 against your order book

Pull your GA4 purchase key event count for the last 30 days. Compare it to your actual order count for the same period. If the numbers differ significantly, your tracking is broken. Every decision on channel mix, ad spend, and product prioritisation is built on bad data.

2. Run your own checkout

Put a low value product through your store as a customer. Watch what arrives in GA4’s DebugView. Does a purchase event appear on the confirmed order page? Does it fire on the payment page before confirmation? Does it fire at all? Does the value match the order and is the currency GBP? This single test tells you more about your tracking setup than any report.

3. Look at your search term reports

If you are running Google Shopping or Search campaigns, download the search terms report for the last 90 days. For Performance Max, the equivalent is under Insights and reports, then Search terms, inside the campaign. Filter by cost. Find the ten highest spend terms. If more than two or three of them would never result in a commercial sale from your core audience, your campaigns are wasting a meaningful share of budget on the wrong traffic.