Ziff Davis VP’s 30-Day Rule for Low-Budget SEO Tests

By James Calloway

Ziff Davis VP of Revenue Jonathan Berthold recommends giving an SEO experiment about 30 days to show a first signal and stopping at 60 days, rather than relying on the old six-to-nine-month advice for new sites.

The longtime strategist has worked across SEO and paid media. That timing is Berthold’s operating rule, not a Google policy or published benchmark. In a Search Engine Journal podcast interview, he treats Google Search Console as directional because, in his view, Google needs time to crawl pages, process changes and adjust rankings.

What should a flat-traffic site test first?

Berthold starts with a full site analysis. He has inherited stalled sites from business-to-consumer software-as-a-service companies and informational lead-generation sites, and the diagnosis determines one of two tests:

• Update existing pages for commercial and transactional queries.
• Build new pages to fill gaps in commercially important topics.

He calls bottom-of-funnel the queries closest to purchase or lead generation. “Top-of-funnel traffic is being siphoned by LLMs, AI Overviews, AI Mode,” he said. The interview included no Google metric for that claim; it was his market observation, not a confirmed ranking or traffic study.

LLMs are large language models; AI Overviews and AI Mode are Google search features. With no budget, Berthold uses AI for research: identify content that ranks for bottom-of-funnel terms, locate competitors’ topic gaps, and find an added perspective that makes a page distinctive to users and Google.

He says terms with more than 1,000 monthly searches often skew informational, where Reddit, LinkedIn and Wikipedia already rank. He calls bottom-of-funnel targeting “100% the easiest low lift opportunity.”

How does he get approval for a test?

Berthold treats approval as communication, not a promise of results. He starts with a hypothesis and states what the test will teach: whether a tactic works, whether a message or creative concept holds up and how the learning can support future earnings.

His pitch connects those lessons to money, which he says leaders operating within a P&L (profit and loss) can evaluate. The request is for a bounded experiment with a decision attached, rather than funding an open-ended sitewide change.

Why start with a small section?

The interviewer raised quick fixes that newcomers use when entering SEO through AEO (answer engine optimization) and GEO (generative engine optimization). Those tactics may work until an algorithm update catches them.

Berthold agrees that a full-site test can tank the organic business. His rule starts with a very small section, then iterates and expands gradually. The SEO equivalent of tripling a paid-media budget in one day, which he says ruins optimization, is rolling out schema (structured-data markup) or new content sitewide and damaging pages that already work.

He also requires replication. A result should repeat often enough to show a pattern rather than a random fluctuation before the team applies it elsewhere.

When should a test stop?

For paid media, where feedback arrives immediately, Berthold reviews results every two weeks or sets a spend limit of two to three times the target CPA (cost per acquisition). Once a test reaches that threshold, the team iterates or kills it.

For SEO, he expects no useful result within a week because crawling, processing and ranking adjustments take time. After 30 days, impressions in Google Search Console usually provide enough directional evidence; he waits no more than 60 days. Starting small keeps expectations controlled and lets other work continue.

What happens after the click?

Berthold sees the same omission in SEO and paid media: teams focus on traffic acquisition and neglect what a visitor does after reaching the site. His candidates are pages already attracting substantial organic traffic or receiving heavy paid spend; they merit split testing and funnel optimization.

He treats landing-page and funnel testing as an always-on discipline. Testing each stage can raise conversion rates even by a few basis points (hundredths of a percentage point), meaning small gains that compound across the path from arrival to conversion.

Can AI citation data guide link building?

Berthold runs an experiment that combines bottom-of-funnel targeting with outreach. He uses AI citation data—references to brands in AI answers—to find publications citing competitors and to spot where LLMs are heading.

Many companies track competitors’ AI visibility without contacting those publications, he says. His logic runs in the opposite direction: a page citing competitors for a bottom-of-funnel keyword likely ranks for that term and appears often in AI answers, so earning a mention can bring direct traffic and place the brand beside competitors.

This remains a prospecting heuristic, not proof that a citation transfers rankings or customers. Berthold calls the approach more holistic because it links content, citations, links and brand visibility.

Is low search volume worth targeting?

Berthold no longer gives search volume much weight when context and topic fit. Google, Moz and other tools may report long-tail terms at one, 10 or zero searches per month, but those figures remain directional estimates rather than a complete measure of demand.

Volume still helps paid media teams estimate the total addressable market, or all potential demand. In organic search, fragmented behavior and LLM query fan-outs make “how your ideal customer profile searches” more useful than a tool’s raw count. His composites-industry example involved fiberglass conduits for commercial projects: if two of 10 monthly searchers reached the site, he said, they could represent a $500,000 contract.

A mortgage-rate searcher likely wants information; “I need a plumber to fix my toilet” signals urgent need and a higher chance of conversion despite low volume. Berthold expects AI usage to make volume even less representative of real queries.

What should marketers monitor next?

His priority is a sequence: find high-intent demand, improve the page where visitors decide whether to convert and measure the business result within one or two months. A conversion number gives leadership a firmer basis for approving the next experiment.

The next useful signal will be whether low-volume, high-intent tests produce repeatable conversion gains, while citation-led outreach adds AI mentions and referral traffic without destabilizing proven pages.