The useful signals are all about method rather than results. Ask what they would do in the first ninety days, who does the work, what you own if you leave, and how they would know the work failed. Vague answers to those four questions predict a bad engagement more reliably than any case study.
Why the market is difficult to read
Search work is bought by people who cannot easily evaluate it, sold by people who all describe themselves the same way, and judged by an outcome that takes months to appear and has many other causes. Those three conditions produce a market where the sales process rewards confidence rather than competence.
It is also a field where the honest answer to most questions is conditional. Someone willing to promise a specific ranking on a specific term by a specific date is telling you something about themselves, not about your site.
Nobody controls the results page. Anyone who says they do is describing a sales script.
The four questions that separate them
What would you do in the first ninety days
A good answer is specific and starts with finding out. Crawl, audit, current rankings, what the site already earns, what the competitors rank for, then a short list of the things that would move most. A weak answer is a package description with a number of blog posts in it.
Who actually does the work
Ask who writes, who makes technical changes, and whether any of it is subcontracted. This is not a moral question. It is a communication question. Work passed through three parties arrives slowly and diluted.
What do I own if we stop
Content, accounts, tracking configuration, and the site itself should all be yours and in your name. Agencies that hold the domain, the analytics property, or the business profile create a hostage situation that has nothing to do with quality of work.
How would you know this failed
The best answer to this is uncomfortable and specific. Someone who has thought about it will name the leading indicators they watch and the point at which they would tell you the approach is not working. Someone who has not will reassure you.
Signals worth taking seriously
- Guaranteed rankings, guaranteed positions, or a guaranteed number of leads.
- A proposal that names deliverables in quantities without saying why those quantities.
- Reporting built around rankings for terms nobody searches.
- Reluctance to show the actual work, or reporting that never contains a page URL.
- A contract with a long minimum term and no defined exit.
- Pressure created by a discount expiring.
A minimum term is not automatically a warning. The work takes months to show, and month to month arrangements can push agencies toward quick visible activity over durable change. What matters is whether the term comes with a stated scope and a way out if the work is not being done.
What changes when the business is local
A local service business is competing on a smaller set of terms against a smaller set of competitors, and the map results often matter more than the organic ones. That changes the work: profile management, service area pages, citations, and reviews carry weight that they do not carry for a national brand.
It also changes the economics. A program built for a company with thousands of pages and a content team is the wrong shape and the wrong price for a business with eleven pages and one person answering the phone. Ask directly whether they work with businesses your size, and what that engagement looks like.
How to actually decide
- Write down what you want to be true in a year, in business terms rather than ranking terms.
- Ask each candidate the four questions and take notes on specificity rather than confidence.
- Ask for two references from businesses of a similar size, and ask those references what went wrong at some point.
- Read the contract for ownership, term, scope, and exit before reading the price.
- Choose the one whose answers you understood.
That last point is not a soft criterion. If you cannot follow the explanation of what someone plans to do, you will not be able to tell whether they did it.
How to choose an SEO agency
- Write down the business outcome you want in a year.
- Ask each candidate what they would do in the first ninety days.
- Ask who does the work and what is subcontracted.
- Confirm you own the content, accounts, and domain if you leave.
- Ask how they would know the work had failed.
- Read the contract for ownership, term, and exit before the price.
Questions people ask
How much should SEO cost for a small local business?
Enough to buy meaningful hours from someone competent, which for most small local programs means a monthly figure rather than a one time fee. Be more suspicious of the cheapest quote than the most expensive one, since low prices are usually funded by volume and automation.
Should I hire a local agency?
It helps for local knowledge and for meeting in person, and it is not decisive. Method matters more than proximity. A specialist elsewhere who understands your market beats a nearby generalist.
What if they will not name specific tactics before I sign?
Some reluctance is fair, since the audit has not happened. Total vagueness is not. They should be able to describe how they work even if they cannot yet say what they will find.
Is a guarantee ever legitimate?
A guarantee about their own conduct can be, such as a defined amount of work or a response time. A guarantee about a search engine's output cannot be, because nobody involved controls it.
How long before I should expect to see something?
Technical fixes can show in weeks. Content and authority work usually takes several months to move rankings, and longer in competitive categories. Anyone promising fast movement on hard terms is describing something else.
What should monthly reporting contain?
What was done with links to it, what changed in traffic and rankings, what is planned next, and what is not working. A dashboard of numbers with no narrative is not reporting.
If you want to see how one shop answers those four questions, the process page describes the first ninety days in detail rather than in packages.
Founder of Big Brain Digital. More than 15 years across SEO, content operations, ecommerce, and AI search. Based in Boise. More about Chris.