Original research

Local SEO studies, run the way studies should be.

Hypotheses written before data is collected. Sample-size math published. Null results published. We're putting our methodology where our recommendations are, and you can audit every number we publish.

Active research

What we're running right now.

One study at a time. Quality over volume; small-N controlled studies with pre-registered methodology beat big-N opinion surveys.

Methodology lockedExpected: Q3-Q4 2026

GBP edit-cadence vs. suspension probability

Controlled study on 60-90 real Google Business Profiles across 3 edit-cadence cohorts. Measures how strongly edit velocity predicts suspension. Pre-registered methodology, full dataset published with results.

Methodology and sample-size math are documented in the repo. Results will publish here once the study runs. If you're an agency or business owner who'd like to participate, get in touch.

The why

Why we run our own research.

Most local SEO advice on the internet, including a lot of our own blog posts, is grounded in two or three canonical studies (Whitespark's Local Search Ranking Factors, BrightLocal's Local Consumer Review Survey, occasional Sterling Sky case studies) and a much larger pile of unverified practitioner opinion.

That's not enough. The local SEO field is small enough that one well-designed study, run honestly and published openly, can shift how the rest of the industry talks about a topic for years. Sterling Sky's 9-business review-count study moved the consensus on the 10-review threshold. We want to publish work that earns that kind of citation.

Our rules: methodology written before data is collected, sample-size math published, null results published anyway, full dataset and analysis code released with every paper. If we say something rose from 16% to 20% in our data, you can pull the data and verify it.

Want to participate in a study?

If you run a local business, an agency, or own a portfolio of Google Business Profiles, we'd value your participation. Real businesses, real data, real impact on the methodology.