Your product feed is your best salesperson.

Your best salesperson never calls in sick. It's a text file. It sits between your inventory and every shopping surface a customer touches. And it works while the lights are off. The desk meets owners who have never opened it. The people buying media above it treat it as plumbing. That's the mistake, and it's the expensive kind.

Treat the feed as staff. You would never put someone on the floor who mumbles the product name, guesses the price and points at an empty stockroom. Your feed does that every day when nobody grooms it. Shopping ads read it. Local Inventory Ads use it to say what's sitting on your shelf right now. Performance Max pulls from it. AI assistants answer shopping questions from whatever they can find about your catalogue. One file, repeated everywhere.

Think about a Saturday afternoon. Your floor staff can properly help one customer at a time. The feed answers everyone at once, in every postal code you ship to, using whatever you last told it. That's an advantage or a liability. The file decides which.

Write the title the way a shopper says it out loud.

Titles do the most work and get the least attention. Brand first, then the product, then the detail that decides the purchase. Colour and size for apparel. Material and dimensions for furniture. Consumer goods usually need the count and the format near the front. Somebody shopping for a sofa types “grey fabric sectional with chaise.” Nobody types “44192 3PC MOD LHF GRY.” One of those titles was written for a shopper. The other was written for a warehouse, and the warehouse is not buying. A title is a shelf label read at arm's length. It has to make sense before anyone decides to look closer.

Price and availability have to be true. Not close. True. Let the price lag your point of sale by a day and the ad carries yesterday's number. The shopper clicks it anyway. Then the number on your site disagrees with the number in the ad. That is a refund request and a one-star review waiting to happen. Availability is worse. An out-of-stock item left live in the feed spends your money sending people to a dead end.

Variants belong in the feed as themselves. Collapse a shoe into one row and you've hidden every size and colour you actually carry. The shopper looking for a nine in black never sees you at all. Furniture tells the same story with fabric and finish. Each real variant is a real search somebody is typing tonight, and a collapsed catalogue quietly opts you out of them.

Images should show the product, not the mood. A clean shot, cropped tight, on a background that doesn't fight it. Lifestyle photography sells beautifully on your own site. Inside a grid of competing thumbnails it reads as clutter. And the grey placeholder box with your logo in it tells the shopper you don't really have the thing.

Categories are the quiet one. Leave the product type blank and something else decides what your product is. A guessed category files your dining chair with office furniture. Then everything built above it learns from the wrong crowd of shoppers. Fill the field in yourself. It costs you one afternoon and stops costing you every week after that.

A broken feed looks exactly like a media problem.

It goes wrong like this. A sync fails overnight. Nobody gets an alert, because the file still uploaded. It just uploaded stale. Two weeks later the numbers are soft and the cost per sale has drifted up. Somebody asks whether the ads need new creative. The creative is fine. So is the bid strategy. Part of your catalogue has been advertising items you can't ship.

So feed health belongs in the Weekly Read, not in a technical appendix nobody opens. When the numbers move, check the data before you touch the media. Disapprovals and availability come before anyone opens the bid settings. A feed doesn't fail loudly. It just stops selling, and the campaign above it takes the blame.

Somebody has to own it by name. Feeds break at the seams between systems, and seams have no owner by default. Your store platform pushes to a channel and a plugin translates the fields on the way. Somewhere in that handoff, a rule rewrites your titles and nobody notices. When the chain is everybody's part-time job, the file rots.

The unglamorous hour is the one that pays.

One hour, the same day each week, one person who owns it. Start with the rejected items, because a blocked product earns nothing at all. Then compare the live count in the feed against the live count in your inventory system. If those two numbers disagree, stop and find out why before you look at anything else.

Then pull your busiest products and read their titles the way a customer would. Out loud is better. You'll hear the ones written for the warehouse immediately. Fix those first, since they carry the most traffic. After that, look at products getting seen and never clicked. That's usually an image or a price, and both are fixable before lunch.

Seasonal work rides on top of the weekly hour. Before a build week, groom the products going into the promotion first. Check that the sale price is actually in the feed and not only on the website. Check that the sizes and finishes you bought deep in are all present and spelled the way people search them. A promotion advertising stock you sold through on day two is worse than no promotion.

Nobody puts feed grooming in a proposal. It doesn't photograph well and it has no rising line on a dashboard. The desk still books that hour before it books anything else. Everything above the feed inherits whatever the feed says, so a title written for the warehouse costs you on every surface at once. That is why we read the file before we argue about budget.

So open your feed this week. Read ten titles aloud and count how many you would actually say to a customer standing in front of you. If the answer bothers you, bring the file to a strategy call and we'll go through it line by line.