The local SEO audit framework I use before I touch any client's Google Business Profile

Most local SEO work starts with an edit. Someone reads a checklist, looks at the profile, and starts changing things. Categories get swapped, descriptions get rewritten, photos get added in a burst, and the team waits to see what happens.

That sequence is backwards. The change you make to a profile is rarely the most important decision in a local SEO engagement. The decision that matters is which changes to make at all, in what order, and based on what evidence. That decision lives upstream of the edit, in the audit.

This article is the audit framework I use before I touch a client’s Google Business Profile or recommend any change to it. It is the same workflow I apply at one location and at fifty. I have run versions of it for storefronts and service-area businesses, single-location operations and franchise systems, in low-scrutiny categories and in the categories Google watches hardest. The structure holds across all of them.

I want to be clear about what this is and what it is not. It is not a 60-point checklist of every signal a GBP can theoretically carry. It is the order I work in, the signals I read first, and the decisions that come out of each layer. The aim is to know what the profile actually needs before you spend any energy editing it.

The companion checklist

Want this framework as a printable PDF to use during actual audits? I built a branded version of the eight-layer audit walk-through, with checkboxes per item, notes space at each layer, and a change-plan template at the end. Download the GBP Audit Checklist (PDF) — no email required.

Why audit first

Three reasons.

Most underperforming profiles have a diagnosis problem, not an effort problem. Teams are willing to do the work. They just do not always know which work matters. Two weeks of new posts on a profile whose primary category is wrong will not move the rankings the team is trying to move. Time and attention are limited; an audit tells you where to spend them.

Editing during a fragile period is its own risk. A profile in a soft suspension, a profile sitting on conflicting data Google has not yet acted on, a profile under a video-verification review: editing during those windows can shift the trust signal in ways you do not want. Knowing which window the profile is in is part of what an audit produces.

You cannot evaluate the impact of changes you have not measured first. A baseline is a precondition for proof. If the team starts changing categories, photos, posts, and review responses at once, and rankings move, you cannot tell which change did the work. Audits give you the before that lets you read the after.

I treat the audit as the work that makes the rest of the work legible. Skipping it is the most common reason local SEO programs feel like they are not producing results when they actually are; the team just cannot tell.

What an audit is reading for

If you read this site you have already seen the brand-trust frame in the reinstatement playbook. The same frame holds for audits. The unit Google is increasingly evaluating is the brand: registrations, domains, user accounts, profiles, directories, the consistency between them. A single profile is one expression of that brand to Google’s systems.

An audit, in practice, is reading the brand-trust picture from the outside. You are not just inspecting the profile. You are reconstructing what Google can see across the registration record, the website, the social profiles, the directories, the user account, the documents, and the profile itself. Where those agree, the brand reads as real. Where they disagree, the disagreement itself is a risk and often the explanation for whatever the team has been struggling to fix.

The audit framework below walks the layers in the order I read them. Each layer answers a specific question. The output of each layer either continues the audit or sends you back to the layer before to fix something foundational before you go further.

Layer 1: Status and account hygiene

This is the layer most teams skip. They start with the visible profile and work outward. I start with the account that owns the profile and the status the profile is sitting in.

What to check

• Is the profile actually suspended or disabled, even if the team thinks it is “live”? Open the Business Profile appeals workflow for the account. The appeals tool surfaces moderation decisions that may not be obvious in the everyday Business Profile Manager view. A profile can be “showing on Maps” and still carry a soft restriction the owner has not noticed.
• Who manages the profile? Pull the owner and manager list. Look for personal email addresses that are no longer in the business, restricted Workspace email addresses, agency or contractor accounts that no longer have a relationship with the business, and accounts that look like they were created just for the profile.
• Is the primary owner account itself healthy? Email-level restrictions, especially on Workspace domains that fail certain checks, are one of the most disruptive things you can carry. If the primary owner is on a restricted account, that account is dragging every profile it touches. Email-level restrictions are difficult to fix or appeal.
• Are there pending edits sitting in the queue? Google can hold edits in moderation for days or longer. If the team has been pushing changes that have not landed, the queue may explain the rankings.

What this layer produces

A status diagnosis: live and clean, live with pending integrity concerns, soft-restricted, hard-suspended, disabled, or account-cascaded.

It also produces a go/no-go for the rest of the audit. If the profile is suspended or disabled, you switch frameworks to the reinstatement playbook and come back to the optimization audit after reinstatement. If the primary owner account is restricted, the next conversation is about replacing the owner before anything else.

I have audited profiles that nobody knew were soft-suspended for months. The team had been making content edits the whole time, wondering why nothing was moving. Knowing the actual status of the profile is non-negotiable as your first layer.

Layer 2: Brand and entity consistency across the web

Once status is read, the audit moves to the brand-trust picture. This is the layer that explains most silent underperformance — profiles that are technically fine but never quite rank where the data says they should.

What to check

• Business registration vs. profile. The name on the government registration record, the name on the GBP, the name on the website footer, and the name on the directories should all be the same. Variations like “Inc.” vs. no designation, “and” vs. ”&”, legal name vs. trade name, are all places I find inconsistency.
• NAP consistency in directories. Walk through the top citation sources for the category. Are the name, address, and phone number consistent across them? Old phone numbers and outdated addresses on aged listings are common.
• The website. Does the footer, contact page, and structured data on the website match the GBP? Multiple addresses on the website when the GBP claims a single location? Multiple phone numbers? Tracking numbers used in ads that have drifted into the directory profile of the business?
• Social profiles. The brand name, the website URL, the address (where applicable) on Facebook, Instagram, LinkedIn, and category-specific platforms. Social profiles that have been left to drift are a brand-trust risk Google can read.
• The business email domain. A business with a website at acmeplumbing.com whose GBP is owned by acmeplumbing.gmail@gmail.com reads as weaker than the same business owned by office@acmeplumbing.com. Domain-matched email accounts are not a published ranking signal, but the brand-trust picture they form is part of what Google evaluates at the brand level.

What this layer produces

A consistency map. Where everything agrees, the brand is reading as one entity to Google’s systems. Where pieces disagree, you have the explanation for either current underperformance or future risk.

Most “ranking” problems I am asked to investigate turn out to be consistency problems with a category dressed up as a ranking problem. The fix is not a content push; it is reconciliation.

Cleaning up the brand-trust picture is foundational work and is often the highest-leverage thing a local SEO program can do. It is also unglamorous, which is why teams underweight it.

Layer 3: Profile classification

Every GBP collapses to one of two classifications: storefront or service-area business. Most teams know this. Most still get it wrong on a meaningful percentage of audits.

What to check

• Does the profile match the actual operation? A profile claiming a storefront with a visible address but no permanent signage at the address is misclassified. A profile claiming service-area only when there is a real storefront customers can visit is also misclassified, just in the opposite direction.
• Coworking and virtual offices. Unless there is a dedicated lockable office with permanent signage and a distinct unit designation, the storefront classification fails the eligibility test. Most coworking and virtual office arrangements do not meet that bar, even when the team has rationalized them otherwise. Google’s Business eligibility and ownership guidelines require that “a business must make in-person contact with customers during its stated hours.”
• Service area definition vs. website service-area pages. If the GBP service area lists 3 cities and the website lists 15, the inconsistency is part of the brand-trust picture. The reverse — website lists 0 service areas, GBP claims a regional service area — is also a problem. The GBP and the website should tell the same operational story.
• For multi-location operations: do the service areas overlap on the map? Two locations of the same brand claiming overlapping service areas reads as a brand-trust problem in my practice. Google has not documented this explicitly, but the pattern is consistent enough across cases I have worked on that I treat it as an operating constraint. Adjacent locations need clean carved territory.

What this layer produces

A classification verdict: storefront, service-area business, or misclassified. The misclassified verdict typically requires fixing before any optimization work begins, because Google’s systems are evaluating the profile against the classification you have declared. If the declaration is wrong, no amount of optimization will perform.

For service-area businesses in trades where Google has elevated scrutiny — locksmiths, garage door, towing, moving, real estate agents, and similar — the classification layer carries more weight than usual. The bar for what reads as a real operation in those categories is higher.

Layer 4: Categories

Primary category is the single highest-leverage optimization choice on a profile. I treat the category audit as its own layer because the decisions here cascade into everything that follows.

What to check

• What is the current primary category, and is it the most specific accurate match? Google’s category list is granular. A “Restaurant” with a more specific category available (Mexican Restaurant, Pizza Restaurant, Bistro, Wine Bar) is leaving relevance on the table. A “Lawyer” with a personal injury, criminal defense, or family law specialization should reflect that in the primary category.
• What primary categories are the top-ranking competitors using for the queries you want to win? This is the cleanest competitor signal you can read. If the top three results for the target query all use a specific category and the profile under audit is using a broader one, the broader category is part of the explanation.
• What additional categories are configured? Per Google’s guidelines: “Use as few categories as possible to describe your overall core business.” Categories should complete the sentence “this business is a,” not “this business has or offers.”
• Are there category options Google has added recently that the profile is not using? Google adds new category options regularly. A category audit two years old may already be stale.

What this layer produces

A category recommendation: confirm current primary, change primary, add or remove additionals. This recommendation is the one I most often see teams skip past on the way to “let’s post more.” Posting more on a profile with the wrong primary category will not produce the results the team is hoping for.

Layer 5: Visible profile data

Now you read what is on the profile itself. This is what every checklist starts with; I get to it fifth.

What to check

• Business name. Exactly the same as the registration, the signage, the website branding, and the citations. Corporate designations (LLC, LTD, Inc, Co) included only if they appear in the real-world brand consistently. Variations are misalignment.
• Address. Format consistent across the web. Suite numbers, unit designations, floor numbers, and abbreviations should all match. Small format variations create partial matches in Google’s systems rather than confident confirmations.
• Phone number. A single primary number across the business. Tracking numbers used carefully, with the brand-primary number as the GBP primary. Old tracking numbers and out-of-service numbers on citations are common drag.
• Hours. Accurate, including holiday hours configured ahead of time. Profiles with stale or perpetually default hours read as less actively managed.
• Description. Within Google’s policy. No prohibited language, no absolute claims, no contradicting service areas, no promotional content that violates Google’s prohibited content rules. The guidelines are explicit about prohibited content categories.
• Website URL. Resolves cleanly, matches the brand domain, no tracking parameters that vary, no broken redirects.
• Attributes. Audited and configured. Attributes drive Maps filtering, and unconfirmed attributes mean the profile does not appear when users filter for them. This layer changes by category as Google adds attribute options.

What this layer produces

A list of profile-data edits, prioritized. Most are small; some are larger. The list rolls forward into the change plan at the end of the audit.

Layer 6: Content layers

Photos, posts, and Q&A. The content layers do real work for the profile, and they are also the layers where most teams overinvest before the foundational layers are clean.

Photos

• Quantity and freshness. Profiles with very few photos underperform profiles with a reasonable library. A working floor is 10 to 15 quality photos, refreshed with regular additions on a monthly or quarterly cadence. The exact cadence is a practitioner observation, not a Google rule.
• Composition. Exterior with signage, interior (where applicable), staff or team, products or work output. Stock-style images and heavily filtered photos drag.
• For service-area businesses without a storefront. Branded vehicle, branded tools or equipment, work-in-progress shots, business card with full branding. The visual evidence is part of the brand-trust picture.
• Format compliance per Google’s photo guidelines: JPG or PNG, 10 KB to 5 MB, in focus, well-lit, minimal alterations.

Posts

• Cadence. Practitioner observation: weekly is healthy, every other week is acceptable, less than monthly starts to look neglected. Google publishes no required cadence.
• Post types per the Posts docs: Updates (general), Events (with required dates), Offers (with required dates and optional coupon code).
• Content quality. A specific headline, two or three sentences of context, a direct CTA to the relevant interior page on the website, a real photo. Generic or stock-feeling posts undercut their own purpose.

Q&A

• Is it being monitored? Q&A is a public surface. Anyone can post a question; anyone can answer. Profiles with unmonitored Q&A often have wrong pricing, wrong hours, or inaccurate service claims sitting there indefinitely.
• Are the most-upvoted answers correct? Practitioner consensus is that the most-upvoted answer surfaces first. Upvoting accurate answers (your own or a customer’s) helps them appear above stale or incorrect ones.
• Proactive Q&A. Pre-answering common customer questions from the verified profile is one of the cleanest small leverage points on a profile.

What this layer produces

A content state report and a cadence plan. The cadence plan is one of the few outputs from the audit that becomes ongoing work for the team rather than a one-time fix.

Layer 7: Reviews

Reviews are simultaneously a social-proof signal, a relevance signal, and a moderation surface. Read all three at once.

What to check

• Quantity, recency, and trajectory. A profile with 30 reviews from two years ago reads as less active than a profile with 30 reviews where the most recent are within the last month. Recency and rate-of-acquisition matter alongside total count.
• Response coverage. Are reviews being responded to, at all, ever, and within a reasonable window? Profiles with consistent owner responses read as actively managed. Boilerplate “Thanks for your feedback” responses are functionally inert; a specific response that references the actual review content reads as real engagement.
• Solicitation practices in the team. Are review requests systematic or one-off? Are they timed against actual service completion? Is the team using a tool or workflow that risks looking like review gating, which is prohibited under Google’s contributed content policy?
• Policy-violating reviews. Reviews that look like incentivized, employment-conflict, off-topic, or fake content are subject to the same contributed content policy as the rest. Flagging review violations is part of profile hygiene.
• What customers actually say. Review content language feeds relevance signals. Reviews that mention specific services, specific cities or neighborhoods, and specific aspects of the business contribute to the profile’s relevance for those terms. You cannot manufacture this. You can prompt customers toward specific feedback by asking what they got help with rather than just whether they were satisfied.

What this layer produces

A review-program assessment: healthy, has hygiene problems, or has solicitation-practice risk. The solicitation-practice risk one is important; a review program that crosses the gating line is one of the surfaces Google can act on, and recovery from a review-related penalty is harder than fixing the practice before it lands you there.

Layer 8: Insights, performance, and behavior

The performance data on the GBP and the linked analytics tell you what the profile is actually doing. Most audits get to this layer and treat it as the entire audit. I treat it as the layer that confirms or invalidates the diagnoses from the layers above.

What to check

• GBP Insights / Performance. Searches, views, calls, direction requests, website visits, message volume. Trend over the last 6 to 12 months.
• Search-term data. What queries is the profile being shown for? Are those the queries the team is trying to win? A profile being shown for the right queries but not converting is a different problem than a profile not being shown for the right queries at all.
• Distance-based performance. For service-area and multi-location businesses, where the searches and conversions are coming from geographically. Pin-rank tracking, where available, adds a layer of resolution.
• Engagement quality. Calls and direction requests are higher-intent signals than profile views. Watch the conversion mix.
• Anomalies. A sudden drop in any metric within the last 90 days is a flag worth tracing. Sudden drops are often the first visible signal of a soft suspension or a policy adjustment in the category.

What this layer produces

Confirmation that the diagnoses from the earlier layers match the data, or a redirect back to a layer that needs re-reading. Performance data should align with the brand-trust picture, the classification, the category fit, and the content state. When it does not, one of the upstream layers is incomplete.

The output: a change plan, not a checklist

The audit ends in a change plan with three properties.

Order. Foundation work first, content work second. If account hygiene is a problem, you do not skip ahead to posts. If primary category is wrong, you do not push more reviews. If brand-trust consistency is broken, you do not start a citation cleanup at the dictionary listings before reconciling the registration record and the website.

Stake-out periods. Some changes need observation periods before the next change layers in. A category change is the cleanest example; switching primary category in the same week as a name change in the same week as a description rewrite makes it impossible to read which change Google’s systems responded to, and it also stacks integrity-check signals in a way that increases risk. Space the high-signal edits.

Measurement. Each high-signal change has a what-we-expect-to-see attached to it. Without an expectation, you have no way to learn from the result. The change plan should record both the change and the prediction.

The output is rarely a 60-item list. It is more often six to ten changes in a sequence, with two or three of them carrying most of the expected impact and the rest being maintenance and cleanup.

When the audit catches a suspension risk before it lands

Audits will sometimes turn up situations where the profile is still live but the brand-trust picture has weakened to the point that a future suspension or category sweep is plausible. This is one of the most valuable outputs of an audit and one of the easiest to dismiss.

The pattern looks like this: NAP inconsistency across major directories, a coworking-style address that does not actually meet the eligibility bar, a primary owner account on a restricted-looking Workspace domain, a service-area definition that contradicts the website, and a category that is more aggressive than the operation supports. None of these alone is necessarily an action. Together they are the brand-trust picture Google’s systems are evaluating, and they tell you a category sweep or an integrity check could land the profile in suspension.

The right move when an audit surfaces this pattern is preventive reconciliation. Move the owner account to a clean address. Reconcile the citations. Switch the classification or fix the storefront eligibility. Tighten the category. Reconcile the service areas with the website. Doing this before a suspension lands is far cheaper than recovering from one.

This is one of the things audits do that checklists cannot. Checklists tell you what to fix; audits tell you what to fix before it becomes the thing the team is panicking about.

A word on what audits do not tell you

I want to be direct about this: an audit cannot tell you precisely how Google’s ranking systems weight specific signals. No one can. The local ranking systems are not publicly documented, and the public statements from Google use general language — relevance, distance, prominence — without quantification. Anyone selling you exact weightings on category fit vs. review velocity vs. citation count is making it up.

What an audit can tell you is which areas of the brand-trust picture are clearly healthy, which are clearly broken, and which are in the middle. With that map, the change plan that comes out of it is grounded in evidence. The team can act with reasoning attached to each decision, can record predictions before changes, and can read results against those predictions afterward.

That is the discipline that makes local SEO programs that compound. The team that audits, plans, edits, measures, and rereads is the team whose results become legible over months instead of remaining a mystery.

The printable checklist

If you want this framework in a format you can actually work from during an audit, the GBP Audit Checklist (PDF) is the eight-layer walk-through with checkboxes per item, notes space at each layer, and a change-plan template at the end. Designed to be printed for in-the-field use, or filled in on screen. No email gate — just download it.

How we help

At Local Visibility Lab the audit framework above is part of how the platform’s brand profile and monitoring features work in practice. Profile-state changes, edit-pattern monitoring, suspension alerts, and rollup reporting at the multi-location level all let agencies and operations teams run audits at scale without manual spot-checking.

If you are managing one profile and want to apply this framework yourself, you can. The layers walk in order and each one produces a question with a real answer. If you are managing 10 or 50 or 200 profiles and need to systematize this work, the platform was built for that.

Either way, the discipline holds: audit first, plan second, edit third, measure fourth. The order matters more than any one tactic.