Why a short list of questions beats a long interview
Most agency calls are won by the person who talks best, not the one who will do the best work. A fixed set of questions fixes that. It makes every seller answer the same things, in the same order, so you can compare them on content rather than charm.
These twelve are written for a store owner who is not an SEO. Each comes with what a good answer sounds like and what a bad one gives away. You do not need to understand every technical term. You need to notice whether the answer is specific or slippery.
One more point before you start. These questions assume you are hiring an agency, but they work for any seller. If you are still choosing between kinds of help, the guide that explains who these questions should go to lays out the options. And if a call goes badly, our page on red flags in ecommerce SEO proposals helps you tell a bad day from a bad fit.
Take notes on the answers, in the seller's own words. When you compare two agencies a week later, you will not remember who promised what.
Questions 1 to 3: who does the work and what goes live
1. Who will actually work on my store, and will I talk to them?
Good answer: names or roles, with how much of their week goes to accounts like yours, and an offer to put you in touch. Bad answer: "our team of experts" with no one named, or a promise that the senior person on the call will be involved, with no detail on how.
2. What will be live on my store by the end of the first 30 days?
Good answer: a concrete list, such as a number of rewritten category pages, fixes made and new pages published. Bad answer: "we will complete an audit and build a strategy." That is a month of planning with nothing changed on your store.
3. Will you make changes directly on my store, or send me a list?
Good answer: they publish directly, with a clear way to review or reverse changes, and they explain which platform access they need. Bad answer: they send recommendations for you or your developer to apply. That puts the hardest part, the doing, back on you. A service that works on your platform directly changes the economics. See what each line of an SEO package produces for how to tell lines that change your store from lines that only describe it.
Questions 4 to 7: content, measurement and AI answers
4. How will the pages you write be different from what every other store publishes?
Good answer: they describe research on your products, your customers' questions and your competitors' gaps, and mention original detail that only your store could supply. Bad answer: "we use proven templates." Template text is exactly what Google and AI assistants see a hundred times.
5. Which numbers will you report, and where do they come from?
Good answer: clicks and impressions from Google Search Console, split by page type, with orders or revenue tied to those pages where your analytics allow. Bad answer: only keyword rankings, traffic totals with no page breakdown, or a score invented by their own tool. Our page on what an ecommerce SEO report should show gives you the model to hold them to.
6. Can you show results from a store like mine, and what was done to get them?
Good answer: a real example with dates, the metric used and the work performed, and honest about what was not achieved. Bad answer: a logo wall, a screenshot with no axis, or a result with no explanation of cause.
7. How do you handle visibility in ChatGPT, Perplexity, Gemini and Claude?
Good answer: they have a way to check whether those assistants name your store and a plan to improve it, such as pages that answer buyer questions plainly and are backed by independent sources. Bad answer: "that is just SEO" with nothing measured, or a guarantee of being named. Nobody controls an assistant's answer. Our guide to how AI assistants pick businesses shows why.
Questions 8 to 10: does it fit your catalog
8. How will you handle filters, variants and out-of-stock products on my store?
Good answer: specific choices. Which filtered pages should be indexed and which not, how variants are grouped into one page, and what happens to pages for products that sell out. Bad answer: a general line about "technical health" with no mention of how your catalog behaves.
9. What do you do about collection pages?
Good answer: they treat collections as the main pages for buyer searches and plan to add real text, FAQs and internal links to them. Bad answer: all effort goes to blog posts and nothing touches category pages.
10. How will you decide what to work on first?
Good answer: a method tied to your data, such as pages with many impressions and few clicks, or categories with strong margins and weak visibility. Bad answer: a fixed order they use for every client, or an answer driven by what is easiest for them to produce.
If you want a neutral reference to hold their answers up against, the complete ecommerce SEO guide covers the full set of jobs a store has.
Questions 11 and 12: cost, exit and ownership
11. What exactly is billed each month, what is one-time, and what are the minimum term and notice period?
Good answer: an itemized quote, a clear term, and a notice period stated without being asked twice. Bad answer: a bundled price with no breakdown, an auto-renewing term buried in the paperwork, or reluctance to put it in writing.
12. If we part ways, what do I keep, and who owns the accounts and content?
Good answer: everything published on your domain stays yours, analytics and Search Console accounts are in your name from day one, and they hand over any documents on request. Bad answer: content lives in their system, accounts are in their name, or exit means losing the pages you paid for.
Two bonus questions are worth asking if the call is going well. First: "What would make you tell me not to hire you?" A serious seller can name a situation, such as a store with no product margin or a site that needs a rebuild first. Second: "What do you need from me each month?" The honest answer is usually small and specific. If the answer is a lot, your team becomes the bottleneck.
How to run the call and compare answers
Send the questions ahead if you can. A seller who answers in writing gives you something to compare later and shows you how they communicate. If they prefer a call, tell them you will ask the same questions of everyone.
Score each answer on three levels: specific, vague, or missing. Do not score on confidence. A specific answer names a number, a page type, a person or a tool. A vague one could be said by any agency about any store. A missing one is a dodge, which is data too.
Pay attention to how they handle "I do not know". Good sellers say it, then explain how they would find out. Weak ones fill the gap with a figure or a guarantee. Be wary of any promise of a particular ranking or a timeline to the top spot, since no seller controls Google or AI assistants.
Finally, ask for a small, real test before a long commitment. A paid first month, a pilot on one category, or a sample of work on your own pages tells you more than any pitch deck. The diagram above shows where to spend your limited time: the top rungs decide most of the outcome.
What to ask their past clients, and what to ask yourself
The best check on any answer is a past client. Ask the seller for two or three stores of a similar size, and ask those owners directly. Keep the questions short, since they are doing you a favor.
What did they actually do for you in the first three months? You are checking the answer to question 2 against reality.
Did you ever have to chase them for a report or an answer? Responsiveness shows up in the first month and tends to get worse, not better.
What would you tell a new client to ask? This question often draws out the one thing the owner wishes they had known.
Would you hire them again? A pause before the answer is information.
If the agency will not connect you with any client, ask why. Confidentiality is possible. A complete absence of anyone willing to speak is not common among sellers who do good work.
Questions to ask yourself
A final group is for you, not them, and it often decides the outcome.
Can I spend the time this needs? Even a hands-off arrangement needs a few hours a month from you to answer product questions and review output.
What would success look like in six months, in numbers I can see myself? Write it down before you sign. A vague goal lets any result be called a win.
What will I do if the first three months show nothing? Decide the exit rule in advance, while you are calm. Search takes time, so a quiet first month is normal. Three quiet months with no visible work on your store is different, and the page on reading each line of a package helps you tell the two apart.
What to do with the answers once you have them
Once you have three sets of answers, put them in a simple grid. List the twelve questions down the side and the sellers across the top. In each cell, write one of three marks: specific, vague or missing. Do it the same day, while the calls are fresh.
Then look at the pattern, not the total. A seller with ten specific answers and two vague ones on cost and exit is worth a follow-up email. A seller with four specific answers and eight vague ones is telling you how the relationship will feel. Missing answers count against a seller more than vague ones, since a vague answer at least tried.
Next, look for answers that disagree with each other. If the seller said on the call that they publish directly on your store and the written proposal says "recommendations will be delivered", the proposal is the real offer. Ask them to reconcile the two in writing.
Finally, weigh fit over polish. The agency with the slickest deck may not be the one that understands a store like yours. Look for the one whose answers mention your kind of problem: your variants, your seasons, your size of catalog. A smaller seller who understands your problem often beats a larger one who understands the average case.
If two sellers tie, choose the one who makes leaving easier. That seller is confident they will not need the contract to keep you.
Adjusting the questions for your kind of store
The twelve questions are general. A few tweaks make them sharper for specific stores.
A store with a large catalog should press on question 8 and ask how the seller handles thousands of near-identical pages without creating thin or repeated ones.
A store with strong seasons should add: when do you start work for a peak that is four months away? Search takes time, so a seller who plans a month ahead will miss the season.
A store that sells specialist or technical products should ask who writes the content and how they learn the subject. Buyers of such goods spot shallow text fast.
A store with a small team should press on question 12 and ask exactly how many hours a month the seller needs from you, and what happens if you cannot supply them.
How RunOctopus handles this
Many of these questions come down to one: will pages go live on my store, and how soon.
RunOctopus is a done-for-you service. Its AI, Ollie, researches your business, then writes and publishes pages directly on your own website. A team plans, writes, reviews and publishes alongside Ollie. It works with Shopify, WordPress, WooCommerce and Wix.
There is nothing to install, nothing to approve, nothing to upload and nothing to schedule. You do not approve each page, because pages go live on their own. In the first week you get 8 long-form articles, 6 pages and a custom interactive tool built for your niche. After that it is 175+ pieces of content a month, written, published and linked. You also get reports on where ChatGPT, Perplexity, Claude and Gemini name your business.
There is a 90-day promise: more visibility on Google and in AI answers within 90 days, measured against the 90 days before you joined, or every dollar back. For scale, an agency doing this work charges about $3,209 a month (2026 average, 439 providers) and publishes 4 to 12 pieces.
Paste your address into the free site read and Ollie reads the site and shows the plan. No account, no call.