Content Lead- Discovery

Zoca; Bengaluru; On-site; INR 900,000 - 1,500,000; Full time

Job summary

Zoca seeks a skilled editor to shape a machine that writes for local businesses. You will define what good content means, prove the machine’s quality, and diagnose failures with rigorous reasoning.

You will own the rubric, weekly reads, and the eval set, aligning content to local intent and measurable outcomes. This role reports into product and collaborates with pipeline engineers to improve performance.

Qualifications

• You have a proven track record ranking local pages for hundreds of small businesses.
• You write clearly and can explain why a paragraph is bad in one sentence.
• You read Search Console to inform opinions.
• You have experience with model-generated content and a measured view.
• You can write a rubric and discuss eval with engineers.
• You have real US local search experience.

Responsibilities

• Define a rubric for good vs bad content and explain why.
• Publish a weekly read and score outcomes across cities and business types.
• Develop an eval set for the pipeline and ensure consistency.
• Define local intent per business type and tailor content accordingly.
• Set and argue about outcome metrics like indexed pages, impressions, and clicks.

Skills

Local SEO

Editorial judgment

SEO best practices

Tools

Google Business Profile

Job description

You will be the editor of a machine that writes for local businesses. You will not write the content. You will decide what good means, prove whether the machine is meeting it, and find the reasons when it is not.

About Zoca

Zoca builds AI agents that local service businesses hire to run the parts of the business the owner cannot get to. More than 1,000 businesses use Zoca today. We started in beauty and wellness and are backed by Accel, Elevation Capital, Better Capital and GTM Fund.

Discovery is the agent that runs a business’s online presence: its Google Business Profile, website, service pages, blogs, Google posts and reviews. It publishes on behalf of thousands of businesses every week, many of them in the same cities.

The terrain

How does content written by a machine, for a business the machine only knows through data, become genuinely useful to a person searching nearby? How does it rank, and keep ranking, when the same machine is writing for the business down the street? Nobody at Zoca owns that question today. Engineers own the pipeline. Nobody owns what comes out of it.

You would inherit a working pipeline from keywords to topics to published pages and posts, a per-business context layer we call Brain that holds services, staff, tone and house rules, a live connection to every business’s Google Business Profile and Search Console, every review those businesses have received, and engineers who want someone to argue with about what the pipeline should produce. Over a thousand live businesses across many cities means a natural experiment running every day, if someone reads it.

What We Know So Far

Google treats content produced at volume to manipulate ranking as spam, whoever wrote it. Fluent, generic content has become free, and specific, local, useful content has not. Many businesses in the same city want the same topics. And from the outside, a page that never got indexed, a page on the wrong topic, and a badly written page all look the same.

What we do not know, and have opinions on that we will share after we hear yours: what separates two businesses’ content in the same zip code beyond the facts about each. How much business-specific data moves ranking compared to the prose. Which content types are worth producing at all. Where a human belongs in the loop and where a human is a bottleneck dressed up as quality. And how you know a page is good before Google tells you two months later.

What You Would Own

• The definition of good. A rubric a person and a model can both apply, with pages that pass, pages that fail, and why.
• The weekly read. A sample of what shipped, across cities and business types, scored against the rubric, with every failure turned into a named class rather than a one-off fix.
• The eval set. The pages the pipeline is measured against before anything goes live, and the automated judge engineering builds from it.
• Local intent, per business type. What someone searching for a lash studio wants to read is not what someone searching for a med spa wants. You decide what we produce for each, and what we stop producing.
• The outcome numbers. Pages indexed, pages earning impressions, keywords ranked, clicks, and calls and bookings we can trace back to a page. Not output volume. Where those numbers should be is yours to argue.

If you find yourself writing blogs, something upstream is broken, and that is the thing to fix.

Who You Probably Are

• You have personally made local pages rank for hundreds of small businesses, at an agency or on a marketplace, and can explain exactly what did it.
• You write well and can say precisely why a paragraph is bad, in one sentence.
• You read Search Console before you form an opinion.
• You have worked with model-generated content and hold a view of it that is neither hype nor contempt.
• You can write a rubric, argue with an engineer about an eval, and lose the argument gracefully when the data says so.
• You have real experience with US local search, wherever you sit.

This is not a content calendar role, a people-management role, a brand copy role or a social media role. It reports into product, works daily with the pipeline engineers, and starts with no team.

Skills: local seo,google business profile,editorial judgment paired with seo