This summer I scored a client’s competitors lever by lever, and one row took me two passes to read correctly.
There were ten competitors in the set, counting the option every buyer has of doing nothing. One of them had published a flat statement that nobody can guarantee what an AI assistant recommends, because the systems keep changing. The first time that competitor went into the record, it was written down as an agency that simply didn’t offer a guarantee, which reads like an empty cell. I corrected it later. An empty cell is what you get from an agency that never thought about guarantees, and this one had thought about them hard enough to write a sentence telling buyers not to trust anyone who offers one.
That sentence turned out to be the most useful thing on their site. It’s probably correct, too, and it still told me more about the buyer and about my client’s opening than their entire services page did.
A competitor’s public refusal is a map: it shows you which outcome buyers want badly enough that someone felt the need to disclaim it, and the exact objection you’ll meet the day you decide to sell it.
Quick Take
- Read what competitors refuse before you read what they claim. Anyone can publish or copy a claim for free, while a refusal only stays on a site as long as the agency keeps turning down the money it costs.
- Competitor refusals come in two kinds. A filter turns away a type of buyer or work. An inoculation play tells the whole market an outcome can’t be had, so whoever promises it next sounds naive.
- You can sell into an inoculation play if your terms are tied to something you control. Promise the exact outcome they disclaimed and you walk into the objection they already planted.
- The openings sit in the commitment levers. Across 53 agency sites I scored, Risk Reversal averaged 2.09, and 44 of the 53 scored a 3 or below.
In this article
- Why a competitor’s refusal is worth more than its claims
- The two kinds of competitor refusal
- How to read an inoculation play
- How to sell an outcome a competitor says can’t be guaranteed
- When a refusal should steer you away
- How to run this on your own competitors
Why is a competitor’s refusal more useful than its claims?
A refusal is the only part of a competitor’s positioning that costs them something to hold. Anything an agency claims about itself can be copied by Thursday, and even the tool advantages behind those claims now expire in months. A refusal only survives on a website while the agency keeps turning down the work it rules out, so when you find one that’s been up for a while, you’re looking at a decision somebody paid for.
I can put a number on how rare that is. When I scored 53 agency websites against the same eleven levers, the four lowest-scoring levers were Relationship Capital, Risk Reversal, Operational Constraints and Economic Model, and every one of them costs the agency something the moment it commits. Operational Constraints, where refusals live, averaged 2.60, and Risk Reversal averaged 2.09. The ranking of the eleven levers came out almost exactly as a ranking of how expensive each one is to mean.
Buyers read refusals this way too. An agency in an RFP-heavy industry put “we don’t participate in RFPs” on its site with no apology attached. A prospect who had just sat through an RFP with three agencies, all sending copy-paste decks telling him what he wanted to hear, called the one agency that refused to play. The refusal did its selling before anyone got on a call. (That story is about an agency’s own refusal, which I cover in how I decide what an agency should refuse. This post is about reading somebody else’s.)
So if a buyer can read your competitor’s refusal and learn something about their judgment, you can read the same sentence and learn something about the market.
What are the two kinds of competitor refusal?
Every competitor refusal is either a filter or an inoculation play, and the two point you in different directions.
A filter turns away a kind of buyer or a kind of work. In one client’s competitive set there’s an agency that serves a single industry and says outright that it won’t take general business clients. That tells you who they’ve decided not to fight for, and it tells you how committed they are to the lane they kept. If your client wants the buyers that agency sends away, the filter is a door. If your client wants the same lane, the filter is a warning that the incumbent has already paid to own it.
An inoculation play is stranger and more useful. The competitor tells the entire market that a certain outcome can’t be promised, by anyone, so the next agency that promises it sounds like it doesn’t understand the work. The AI recommendation line from my client’s set is one, and it’s far from unusual. Sprocket Digital, an agency in New Zealand, writes on its blog that a reputable agency “can’t and won’t make guarantees simply because it’s near impossible to do so,” and tells readers to “run for the hills” if an agency guarantees results at the start. AudienceIntent, which sells AI visibility work, published a piece in July 2026 headlined “No One Can Guarantee You’ll Be the Number 1 Answer in AI Search.”
Neither of those is dishonest, and both are probably right about the outcome they describe. They’re also doing a sales job, because every buyer who reads them now carries a ready-made reason to distrust your guarantee before you’ve made one.
How do you read an inoculation play?
Read it for the demand it reveals, the objection it plants, and the gap it leaves behind.
Start with demand. Nobody writes a paragraph disclaiming an outcome buyers don’t ask for. If a competitor felt the need to tell a market that AI recommendations can’t be guaranteed, buyers in that market are asking for exactly that.
Then write the objection down, word for word. When your client eventually offers some kind of assurance in that market, a prospect who has read the competitor will say a version of that sentence back on the call. You get to prepare for it weeks in advance.
The gap matters most. In my client’s set, Risk Reversal came back None or Weak for all ten competitors, and the only de-risking anywhere was one vendor’s free tier. The refusal and the empty column feed each other. If nobody in a market will carry any risk, and one of them is telling buyers nobody should, the safest-looking choice is to do nothing, and doing nothing was already the default option in almost every one of those deals. One owner in that market had cycled through website managers for eight years. For a buyer like that, waiting reads as prudence.
The same shape showed up for a different client in a different industry, where Economic Model and Risk Reversal came back empty across all four of its main rivals. Both of those sit in the commitment bucket of the eleven levers, the bucket where most agencies are completely empty, so it’s also the most available place to build separation.
How do you sell an outcome a competitor says can’t be guaranteed?
Agree with the refusal in writing, then guarantee something your client controls. Promising the exact outcome the competitor disclaimed puts your client inside the objection that competitor planted, and the prospect will be right to push back.
Guarantees that hold up attach to inputs and conditions the agency owns. Mykyta Hryhorenko at Gritraffic put it plainly in July: “An honest guarantee attaches itself to something the agency actually controls.” His other line is worth taping somewhere you’ll see it: “the thinner the actual protection, the louder the language tends to get.”
When I help an owner build one, it gets designed in three layers. The first is a trigger somebody can check, like “the system is running by day 90,” where most agencies would write “you’re satisfied.” The second is what it costs the agency when it fires, whether that’s a refund, free continuation, or the owner’s own time until the thing works. The third is what the client has to do to collect, and that condition is what stops a guarantee from attracting people who want a refund more than a result.
One coach I worked with designed his live on a working session. He refused to offer money back, because in his words “you start attracting people who really want their money back and aren’t actually serious about the goal.” He built a completion guarantee: finish every check-in and implementation on time and to standard, and if the system still isn’t in place, he works with you personally until it is. He paired it with a 30-day parachute, where a client who’s unhappy in the first month simply doesn’t get the next invoice.
Now hold that up against the inoculation play. The competitor said nobody can guarantee what an AI assistant will recommend. Fine, say so on your own page. Then guarantee the part your client runs, with a checkable trigger, a cost to the agency, and a condition the client meets to collect. In a set where nobody carries any risk, your client becomes the one agency that does, and the competitor’s own sentence makes the offer read as more honest.
One condition before any of this. A guarantee is only cheap to offer when the outcome pattern behind it already exists, because terms like these expose inconsistent results fast. If you can’t yet show the same result more than once, build that first.
When should a competitor’s refusal steer you away?
When the thing they refuse is something your client couldn’t carry either. Some refusals are just correct. Guaranteeing a number you’ve never measured will come back to bite you, and taking on downside without a repeatable delivery system is gambling with extra steps.
Copying the pack is the other trap. Every lead gen agency says “you don’t pay until you get a lead,” and the line stopped meaning anything once everyone said it. In a set where every competitor already offers the same promise, a refusal from one of them could be the sharpest position in the room, and adding your client’s name to the pile helps no one.
And when a competitor refuses a kind of client that tends to wreck agencies, believe them, because they paid to learn it. Every refusal is a trade somebody decided to make, and the ones about bad-fit clients are the cheapest lessons you’ll ever get secondhand.
How do you run a refusal read on your own competitors?
Give it an hour, and do it before your next homepage rewrite.
- Pick three competitors a buyer would realistically compare you to, and add a fourth column for doing nothing, since that’s the option a stalled deal defaults to.
- Go through each site and pull every sentence that refuses something: “we don’t,” “we only,” “we won’t,” “no one can.”
- Sort each one into a filter or an inoculation play.
- For each filter, decide whether the buyers they’re turning away are buyers you want.
- For each inoculation play, write down the objection it plants, then work out a checkable trigger you control, what it would cost you when it fires, and what the client does to collect.
If you finish with nothing worth offering, that’s a finding too. It means the levers worth working on are somewhere other than risk, at least for now.
If reading competitor sites usually leaves you feeling worse about your own, you’re in good company. Reading them this way tends to do the opposite, because every refusal you find marks a spot they’ve told you they won’t go.
If you want a second set of eyes on your first one, bring it to the Dynamic Agency Community.
FAQ
Should my agency offer a guarantee if competitors say results can’t be guaranteed?
Often, yes, as long as it covers something you control. Agree with them that nobody controls a platform’s algorithm, then guarantee a checkable trigger inside your own delivery, name what it costs you when it fires, and name what the client does to collect.
Is it a red flag when a marketing agency guarantees results?
It can be. A vague promise of “results” with no metric, no deadline and no way to check deserves the skepticism it gets. A guarantee tied to something the agency runs, with the terms written on the page, deserves to be judged on those terms.
How do I find what my competitors refuse to do?
Search their sites for “we don’t,” “we only,” “we won’t,” and “no one can,” then open the FAQ. One agency in my 53-site sample had a stated refusal tucked inside a collapsed FAQ accordion at the bottom of its homepage.
What if none of my competitors refuse anything?
Then the commitment levers are probably wide open. Risk Reversal, Operational Constraints and Economic Model were three of the four lowest-scoring levers across the 53 agencies I scored, so a competitive set with no refusals and no guarantees is normal, and the first agency in it to hold a line stands out.
Can a competitor copy my guarantee once I publish it?
The words, yes, and quickly. The delivery system that makes the terms affordable takes much longer to build. Without it, a copied guarantee usually grows fine print or costs the copier money.