Citation building: what is worth doing and what is theatre

Citations are the most billable thing in local search. A vendor can deliver two hundred of them in a week, screenshot the list, and produce a report that looks like a month of labor.

The customer cannot evaluate any of it. Nobody checks whether a listing on a directory they have never heard of did anything, because there is no way to check.

What is a citation, exactly?

A mention of your business details somewhere other than your own site.

Two kinds exist. Structured citations are directory entries with fields: name, address, phone, hours, category. Think Apple Maps, Yelp, Bing Places, and the industry directories in your field. Unstructured citations are mentions inside ordinary content, a local news article, a neighborhood blog, a sponsor list on a school fundraiser page.

The original logic was that Google used these mentions to confirm you exist and to build a picture of how established you are. That logic still holds, and it is the reason the core set is worth doing.

What has changed is the volume required, and the direction of travel is documented. Whitespark's 2026 Local Search Ranking Factors report, in which 47 local search experts scored 187 factors, has HTML NAP matching your profile NAP at number fifteen with 153 points, inside the top twenty but well below category, proximity, hours, and reviews. Darren Shaw's own summary of what shifted in the 2026 edition is that review and behavioural signals rose while citations lost about half a percentage point, continuing a downward trend.

Read that carefully, because it is not "citations do nothing." It is "consistency still counts and volume never did." Confirmation does not need three hundred sources. It needs a handful of credible ones that agree with each other.

Which citations are actually worth building?

The ones a real customer might plausibly use, plus the ones specific to your industry.

For almost every local business that means the same short list: Google, Apple Maps, Bing Places, Facebook, Yelp, and the main data aggregators that feed smaller directories. Then the industry set, which is where the real value sits. A contractor belongs on the trade directories customers actually search. A dentist belongs on the insurance and provider directories. A restaurant belongs on the reservation and delivery platforms.

A citation on a site your customers use is worth more than fifty on sites nobody visits, because it can send you a customer directly. That is the test. Would a human being ever arrive at that page on purpose?

There is one piece of research worth knowing when you pick the review-bearing ones. Whitespark's Miriam Ellis ran 153 queries across 17 local business categories in 9 US cities and documented the review sources attached to each listing on the first page of Bing Places. Facebook was the leading review source, appearing on about 1.5 times as many businesses as Yelp. Yelp came second with more than twice the occurrences of third-place TripAdvisor, followed by Porch, YellowPages and Angi. Category variance was large: TripAdvisor dominated bakeries and sushi, Porch dominated garage door repair.

Two caveats travel with that. It is a Bing Places proxy, not a direct measurement of any AI assistant, and the study's own methodology section says 17 categories while its findings reference 18. Whitespark sells local SEO software. Use it to pick which three or four platforms to bother with in your category, not as a ranking law.

Chamber of commerce pages, local business associations, and neighborhood organizations are quietly good. They are real, they are local, and they are usually cheap.

Sites that matter for reviews deserve extra care, because each one has its own solicitation policy and Yelp's rules differ sharply from Google's.

Why are bulk directory packages theatre?

Because they optimize for a countable number rather than an outcome.

The listings are on sites with no traffic, often auto-generated, frequently duplicated across networks owned by the same company. Nobody visits them. They send no customers. Their contribution to how Google sees you is, at best, small and, per the survey above, shrinking.

The best available number on where local calls actually originate makes the point harder than any argument about ranking weight. Patient Prism's Dental Patient Access Report, published July 2026 and built from 11,552,668 patient calls across 8,280 locations with source attribution on 5,022,887 of them, put 90% of attributable calls inside the Google ecosystem. The Business Profile alone was 54%, organic and Google Ads another 36%. Meta and print combined came in under 3%. Everything else in the world, every directory a bulk package would sell you, competes for what is left of the remaining tenth.

Grade it: that is one vertical, dental, and it is a call-tracking vendor reporting on its own customer base, which skews toward practices that already invest in this channel. The proportions would move for a restaurant or a plumber. They would not move enough to make a three hundred pack worth buying.

They also create a maintenance liability. Every listing is another place your details can go stale after you move or change a number, and the volume makes cleanup genuinely painful later. You bought two hundred future problems.

And many are rented. The listings persist while you pay, and can revert when you stop. Ask that directly before signing: what happens to these listings if I cancel in six months?

What does a three hundred listing package actually cost?

Here is the arithmetic. Substitute your own quote.

The invoice. Say the package is $199 a month for 300 listings, managed. That is $2,388 a year, renewing.

The alternative. The core set for most businesses is eight to ten entries. At fifteen minutes each that is two to two and a half hours, once, plus twenty minutes a quarter to confirm nothing drifted. Call it three and a half hours in year one.

Now price the exit, which is the part nobody quotes. You change your phone number in March. Updating eight listings takes two hours. Updating 300 at four minutes each is 20 hours. You will not do 20 hours. So the old number lives on 292 pages indefinitely, and the consistency signal the package was sold to build is now actively working against you.

Then price the upside. In the Patient Prism data, the entire non-Google world accounted for the remaining 10% of attributed calls, with Meta and print inside that at under 3%. Split the difference generously and assume your bulk directories are worth a percentage point. You are paying $2,388 a year for a slice of one percentage point of your call volume, and you cannot measure which slice.

The honest comparison. Take the same $2,388 and put it against review cadence, a faster phone answer, or three genuinely local unstructured mentions. Every one of those has a shorter path to a customer.

Do citations still matter for AI answers?

The unstructured ones do, and that inverts the usual sales pitch.

Whitespark's May 2026 testing of AI Mode local answers found that star ratings, review volumes, pricing and hours are pulled from Business Profiles or other sources, and that reviews and unstructured citations, meaning blogs, news and community pages, are the heaviest inputs to AI Mode local answers. That is practitioner testing rather than a controlled experiment, so hold it loosely. But note what it says: the neighborhood blog writeup and the local news mention, the two things a bulk package cannot manufacture, are the ones with a live role.

The one cheap hedge worth taking is Bing Places. Which index ChatGPT uses for local is genuinely contested, Whitespark's own conclusion is that there is no firm industry consensus, and claiming Bing Places costs an hour with no downside.

That is roughly the whole honest answer on AI and local, and it is close to what I actually tell a client who asks about AI.

How do I know if citation work is doing anything?

You mostly cannot, and that is the point that gets skipped.

Citations are a confirmation signal. When they are missing or contradictory, fixing them helps. When they are already fine, adding more produces nothing you can measure. There is no clean before and after, which is exactly why the category is easy to sell and hard to hold to account.

Nobody publishes a controlled test showing that citation count moves local rankings, and the one large recurring measurement in the field, the Whitespark survey, records the signal declining year over year. Treat any vendor claim of a citation-to-ranking relationship as an assertion, and ask what it was measured against.

So judge the work by whether it changed something specific: a wrong phone number corrected, a duplicate listing removed, a missing industry directory added. Those are real. "Two hundred citations built" is not a result, it is an activity.

This is a good test for any local report you receive. A monthly report should tell you what changed and what it cost, not how many tasks were performed.

What to do this week

List the five platforms your customers in your industry actually use. Not the platforms an agency listed. The ones you would name if a friend asked where to look you up.

Check each one. Is the listing there, claimed, and correct? Fix what is wrong, claim what is unclaimed, and add what is missing. For most businesses this is two hours, once.

Claim Bing Places while you are at it, if you have not. One hour, no downside.

Then spend the rest of the time on something with a shorter path to a customer. Seeding your profile's Q&A is a good candidate, since it sits directly in front of people already deciding, and there is a right way to seed it.

If a competitor is outranking you and you suspect their listing is not playing straight, that is a different job with a real process, and a keyword-stuffed business name can be reported.

Questions about a citation package you were quoted?

Email me at eric@seod.com with the proposal, or just the list of sites it promises. I will tell you which entries are worth having, which are filler, and what I would pay for the useful part. If the whole thing is padding, I will say so and you can keep the money.

I read these myself. No sequence follows, and I have no citation product to sell you instead.

The core listings are worth getting right. Chasing hundreds of low-value directories usually is not. Search visibility treats citations and NAP consistency as a foundation rather than a standalone add-on. Or keep reading the rest of the local search library.