Feedcast: where we come from, where we're going

We're starting a series of articles to tell you what we're building, as we build it. Might as well start at the beginning: where Feedcast comes from, and where it's headed.

It started with a very concrete problem

Feedcast was born out of a frustration every online retailer knows: your product catalog is your main commercial asset, and nobody really helps you keep it alive on the channels where your customers actually buy. Google Shopping, Meta, Microsoft, marketplaces — each one has its own requirements, formats and rejection reasons. The work is thankless, invisible and never finished.

So that's where we started: connect a catalog, distribute it cleanly everywhere, and run the campaigns behind it. With one structural advantage we earned in 2020 and have held ever since — Google CSS Premium Partner status, which makes your Shopping ads up to 20% cheaper per click than serving them directly through Google.

Today, more than 6,300 stores are connected to Feedcast, and the average ROAS observed across our base is 4.2×. That's the foundation. It isn't the destination.

What changed over the past two years

AI made economically feasible what used to be absurd: properly rewriting tens of thousands of product listings. So we added automatic enrichment of titles and descriptions, then attribute extraction (brand, color, material, size) straight from your existing content.

The real turning point was somewhere else, though. After automating tasks one by one, we realized the hard part wasn't execution — it was the decision. Knowing what to fix first, which product to push, which price to adjust, which campaign to cut. That's what eats your time, and that's what nobody does for you.

Where we're going: a growth agent

Our conviction is that the next step isn't one more tool to operate, but an agent you delegate to. An agent wired directly into your catalog, your campaigns and your prices, working while you do something else, and reporting back in the morning.

The shift is already under way market-wide: according to eMarketer, roughly $20.6 billion of US retail sales will flow through AI platforms in 2026 — nearly four times more than in 2025, or 1.5% of US e-commerce. Shoppers are starting to delegate their research to agents. Merchants will delegate their operations.

Our line is simple, and we stick to it: the agent recommends, you decide what you hand over. We move one step at a time, and you keep control over every additional notch of autonomy. Trust isn't declared, it's earned.

What you'll read here

This series is our way of working out loud. Concretely, you'll find:

• product news, explained with the reasoning behind it — not just a feature list;
• what we learn about catalog quality, ad performance and pricing, based on what we observe across our base;
• our methodology choices, including when they change — like the FeedScore rework we cover in the next article.

No rigid schedule: we publish when there's something useful to say. Important releases will also reach you by email, and we'll relay the essentials on our social channels.

A question, a disagreement, an idea? The chat is open in the app, and we read everything.