Get a Feed Readiness Review
ChatGPT Ads & AI shopping placements

Your Google Shopping feed is not enough for ChatGPT Ads.

Shoppers do not type keywords into ChatGPT. They describe what they need. Lead Horse restructures your product feed so AI shopping systems can understand what your products are, when to show them, and why someone should click.

Built for ecommerce teams preparing for ChatGPT Ads
Optimized around real buyer intent, not just product names
Feed strategy, content, variants, image context, eligibility, and QA
feed_row · as exported

Toggle the card. Same product, two very different feeds.

The real problem

People do not ask ChatGPT for SKUs. They describe what they need.

A shopper may ask, "What is a good tote bag that can also be worn as a backpack?" If your feed only says "25L Convertible Carryall," the system has to connect the dots on its own. Give it the words, the use case, the color detail, the Q&A, and images that prove both carry modes, and you give it reasons to match your product to the right shopper.

AI shoppinglive intent
What is a good tote bag that can also be worn as a backpack for commuting?
A few that fit that
Convertible 25L Tote Backpacktote + backpack · 25L · laptop sleeve
recommended
DayHaul Convertible Carryhands-free · commute · 22L
recommended
Your bag"25L Convertible Carryall" · no carry modes, no use case
not shown
Same product. The feed is the only thing that changed.
Shoppers ask
"What is a good tote bag that can also be worn as a backpack?"
"What is a durable bottle for work and commuting?"
"What are practical gifts for someone who travels?"
Weak feeds say
25L Carryall  ·  22 oz Bottle  ·  Holiday Bundle
Stronger feeds explain
Convertible tote backpack with hands-free carry
Insulated stainless steel bottle for work, travel, and daily hydration
Gift-ready travel essentials with clear product context
The fields that decide performance

The feed fields most brands do not know they need.

Title, description, price, availability, product URL, image URL. That is enough to get a feed started. It is not enough to make a product easy to understand in ChatGPT Ads or AI shopping placements. These channels use fields and structure most Google Shopping feeds were never built around.

01 / Eligibility fields

Not every product should be eligible for everything.

Your feed may need to define search eligibility, ads eligibility, and checkout eligibility per product. Checkout should stay off until fulfillment, policies, support, and order handling are ready.

Most brands treat eligibility like a technical setting. It is product selection, budget protection, and launch strategy built into the feed.

eligibility
searchtrue
adstrue
checkoutfalse
checkout stays false until fulfillment, policies, and support are ready for direct checkout
02 / Product Q&A

Answer the questions shoppers actually ask.

Not just "What is the capacity?" but the real buying questions that decide a sale. This gives the system more context around when your product should be recommended.

product_qa
Is this good for commuting?
Can this be carried as both a tote and a backpack?
What can fit inside?
Is this better than a regular tote?
03 / Variant dictionaries

"Gray Multi" does not explain what the shopper sees.

It may be technically correct, but a stronger variant structure describes the product the way a person looking at it would. That is more useful for product understanding, shopper expectations, and variant matching.

variant_dictionary
colorGray Multi
display_colorGray Pink Multi
primary_colorGray/Taupe
accent_colorsPink, Light Blue
04 / Use-case metadata

Name the intent, not just the product.

Your feed can clarify the buying intent behind a product: what it is, how it should be positioned, and which shopper need it should match.

This is not keyword stuffing. It is structured context.

metadata
intentwork tote backpack
winning_termsconvertible tote, tote backpack
seasonalityevergreen
product_focus25L Convertible Carryall
05 / Image context

If the title says convertible, the images should prove it.

Images help explain the product. The set should show tote mode, backpack mode, interior capacity, on-body scale, detail shots, and lifestyle use. Image order, additional image URLs, and even file naming all support product understanding.

image_set · ordered
1front.jpg
2backpack-mode.jpg
3tote-mode.jpg
4interior-capacity.jpg
5on-body-scale.jpg
weakIMG_4827.jpg
betterblack-green-carryall-backpack-mode.jpg
06 / JSON readiness

A feed can look clean in a spreadsheet and still break.

Fields like Q&A, variant dictionaries, metadata, reviews, geo pricing, and availability may need to become real JSON objects, arrays, and booleans. A feed that looks organized is not always ready to submit.

feed.json
{
  "id": "ubp-25l-graypink",
  "eligibility": { "ads": true, "checkout": false },
  "accent_colors": ["pink", "light blue"]
}
booleans as textarrays as stringssmart quotesduplicate IDsmalformed image listsblank fields
Valid is not the same as ready

A technically valid feed can still be a bad feed.

Most brands stop when the feed uploads without errors. That is not enough. A feed can have valid product IDs, prices, URLs, images, and availability and still fail to explain the product in a way AI shopping systems can use.

Technically valid

  • Product ID exists
  • Price is formatted
  • Product URL works
  • Image URL works
  • Availability is in stock
  • Title is not blank

Actually useful

  • Title matches buyer language
  • Description explains the use case
  • Variants are grouped correctly
  • Colors match what shoppers see
  • Q&A answers real shopping questions
  • Images prove the product claim
  • Eligibility is set intentionally
  • JSON converts cleanly
  • Product selection is focused

Technical validity gets the feed accepted. Product understanding gives it a chance to perform.

If the feed cannot explain the product, the system cannot recommend it.

We find every gap between how your buyers ask and what your feed says.

Get a Feed Readiness Review
Before and after

The product did not change. The feed did.

Before
Title
25L Convertible Carryall
Description
Our bestselling carryall is perfect for everyday adventures.
Color
Gray Multi
Image
gray-front.jpg
The problem
  • Too vague. Does not explain that it is a tote backpack or who it is for.
  • No reason to match it to work, commute, travel, or hands-free queries.
After
Title
Work Tote Backpack for Laptop and Commute, Gray Pink Multi
Description
A roomy work tote backpack for commutes, office days, and carrying more without switching bags. Use it as a tote for quick access or wear it as a backpack when your load gets heavier.
Variant data
Gray/taupe body with pink and light blue accents.
Image set
front, side, inside, tote mode, backpack mode, on-body scale.
Q&A
Is this a good work bag? Can it be carried as a tote and a backpack? What can fit inside? Is this good for commuting?
Why it works
  • Clear category, use case, color data, and image context
  • Eligibility set intentionally, with a better match to natural-language intent
What Lead Horse does

We turn your product feed into an acquisition asset.

Lead Horse does more than clean up columns. We connect product data to buyer intent, feed structure, paid media strategy, image context, and launch priorities.

01

Pick the right products to launch first

We help decide which products should be eligible based on demand, margin, seasonality, inventory, product clarity, image quality, and landing page strength.

02

Map products to buyer intent

We look at how customers actually describe what they need, then align titles, descriptions, Q&A, images, and metadata to that intent.

03

Rewrite titles and descriptions

We replace vague internal product names with clear, intent-driven feed content built around how shoppers ask for products.

04

Fix variants and color structure

We clean up item IDs, group IDs, item group titles, variant dictionaries, display colors, primary colors, accent colors, and color descriptions.

iditem_group_idvariant_dictionarydisplay_coloraccent_colors
05

Improve image strategy

We review main images, additional images, image order, image naming, lifestyle coverage, and whether the image set proves the product claim.

06

Set eligibility intentionally

We help decide which products should be eligible for search, ads, and checkout, and which should stay out of the launch feed.

search_eligibilityads_eligibilitycheckout_eligibility
07

Validate and prepare for JSON

We check nested fields, arrays, booleans, required fields, empty values, product URLs, image URLs, price, availability, and upload risk.

Feed readiness score

Is your feed ready, or just uploadable?

Answer honestly. The score is a self-check, not an audit, but it shows where most feeds quietly lose performance before launch.

0
Feed readiness score
Answer the questions
Most feeds score lower than their owners expect.
Get a Feed Readiness Review

If you cannot answer yes to most of these, fix the feed before you fund the campaign.

What you get

An audit, a feed rewrite, a launch-ready build, or a developer handoff.

You can use us for a focused audit, a feed rewrite, a launch-ready feed buildout, or a developer-ready handoff. Every engagement is scoped to what your feed actually needs.

Who it is for

Built for ecommerce brands launching into AI shopping channels.

This is for you if

  • You are preparing for ChatGPT Ads
  • You already have a Shopify, Google Shopping, or Meta catalog feed
  • You have strong products but messy product data
  • You need to decide what products to launch first
  • Your feed has fields your team does not know how to use
  • Your image set does not prove the product claim
  • You want marketing, ecommerce, and development aligned before launch

This is not for you if

  • You want to dump your full catalog with no strategy
  • You do not know your margins or inventory
  • Your product pages are not ready for traffic
  • You are unwilling to rewrite feed titles or descriptions
  • You want a one-click export with no optimization
  • You want to mark every product eligible for ads
Feed readiness review

Get a Feed Readiness Review.

Send us your feed, website, and the product or category you want to launch first. We will show you what is technically broken, what is strategically weak, and what needs to change before you spend.

No generic audit. No bloated report.
A practical review of what needs to change before launch.
Reviewed by the team that does the work, not a sales rep.
1 · Your details 2 · Prep for the call

Takes 20 seconds. We will follow up at this email, and the next step helps us prep for your call.

Your details are in. We are on it.

We have your details and your answers. We will review your setup and reply within two business days to book a call and walk through the biggest risks.

Questions

What brands ask before they start.

Not always. You may be able to start with your existing Shopify or Google Shopping feed, but it usually needs to be adapted. Titles, descriptions, variants, images, Q&A, eligibility fields, and JSON structure often need work before launch.

Google feeds are a starting point, but AI shopping systems need more product understanding. Fields like Q&A, variant dictionaries, display colors, accent colors, image order, image naming, and eligibility settings can affect how well the product is understood and matched.

Usually, no. Start with the products most likely to convert. A smaller, cleaner launch feed is often better than a broad catalog dump.

The image itself matters most, but image names and URLs can support product understanding. A file named backpack-mode.jpg gives more context than IMG_4827.jpg.

Yes. We rewrite titles and descriptions around buyer intent, product function, and use case, not just internal product names.

Use-case Q&A answers shopping questions like "Is this good for commuting?" or "Can it be carried as a backpack?" It helps the feed explain when the product should be recommended.

Yes. We look at demand, margin, inventory, seasonality, image quality, landing page strength, and product clarity to recommend where to start.

We can audit the feed, clean the CSV, rewrite product fields, prepare JSON-ready data, create a validation report, or hand off instructions to your developer.

Yes, if needed. But the first step is making sure the feed is strong enough to support the campaign.

Last thing

Fix the feed before you fund the campaign.

ChatGPT Ads can only work with the product data you give them. Make sure your feed is clear, structured, and built around the way your customers actually shop.