Section 1
Three shortlisted providers, one identical strip of logos
Picture the shortlist. A 60-seat law firm needs a new IT provider. The office manager, who is not technical and did not ask for this job, has three proposals open in three tabs. All three sites carry a strip of logos near the footer. A Microsoft partner designation on all three. A security vendor badge on two. A backup vendor badge on two, though not the same two. One firm has photographed a framed certificate hanging in its office. Ask what that strip does for the office manager. She cannot tell whether one Microsoft designation is harder to earn than another. She does not know which vendor badges are audited and which arrive with a reseller agreement and a minimum spend. She has no way to connect any of it to the thing she actually fears, which is calling at 4pm on a Friday with the practice management system down and getting a ticket number instead of a person. So she does the rational thing. She treats all three as equal on credentials and moves to a column she can read, which is the monthly number. Every badge on that page just contributed to a tie, and ties get broken on price. That is the mechanism, and it is worth stating plainly. It is not that your certifications are worthless. It is that a proof only separates you from a weaker provider if the weaker provider could not produce the same proof. Every firm in your metro can put those logos on a page by next Tuesday, and the buyer cannot rank them anyway. The real question is less comfortable than it first looks: which of the things on my website could the worst MSP in this city not copy? Answer that honestly and most of your marketing decisions make themselves. The rest of this is how to find those things, and how to be fair to certifications where they genuinely earn their place.
Section 2
The honest case for certifications, made properly
Before arguing against certifications as marketing, it is worth making their case better than most of their critics do. Three different things get lumped under one word, and they behave differently. Individual technical certifications sit with the engineer. They evidence that a person worked through a syllabus and passed an exam, and they walk out of the door when that person does. Vendor partner tiers sit with the firm and are usually earned through some mix of exam passes, revenue and customer count. They unlock economics: discount bands, deal registration, not-for-resale licences, marketing funds, and often a support escalation path that skips the public queue. Firm-level audited standards are a different animal. SOC 2, ISO 27001, CMMC and their equivalents mean an outside party examined how you actually operate. They cost real money and real months, which is why they carry more weight. Now the places where these plainly do sell work. Compliance-driven buyers. A defence subcontractor, a clinic, a firm handling card data. Their own obligations flow down to their vendors, so the certification is not a marketing claim, it is a procurement input. Insurance and enterprise procurement. Cyber insurance questionnaires and enterprise vendor forms have fields. An empty field can end the conversation before anyone reads your proposal. Vendor economics. Tier status affects your margin directly. If a higher tier moves your licensing cost by a few points across your whole base, that is a real number, and it has nothing to do with your website. Escalation access. A partner tier that gets you a named contact when a platform is broken is an operating asset your client will feel, even though they will never see the badge. None of that is in dispute. The claim here is narrower. Those are procurement gates, cost structures and operating capabilities, not differentiation in a competitive shortlist where the buyer is non-technical and your rivals hold similar credentials. Confusing the two is how owners end up investing in badges and wondering why the close rate never moves.
Section 3
The separating test: three questions to run on any proof
A signal only carries information when it is expensive to fake. That is the whole idea. If a strong provider and a weak provider can both display something at the same cost, the display tells the buyer nothing about which one they are talking to. Run any piece of proof on your website through three questions. Could my worst local competitor have this by next month? If yes, it is table stakes. Vendor badges, most partner tiers, a certificate on the wall, a testimonials page you assembled yourself, awards with an entry fee. None of these are lies. They just do not sort. Can the buyer rank it without asking me? A non-technical buyer cannot order two partner designations by difficulty. If ranking requires your explanation, the proof does no work at the exact moment it matters most, which is when you are not in the room. Does it cost me anything if I fail to deliver? This is the question that does the separating. A claim with no consequence attached is free to make, so everyone makes it. A claim with a consequence attached is only affordable to a firm that can actually meet it. There is an asymmetry worth naming, because it explains why badges persist and why you should keep yours. Table stakes are not neutral. Their presence adds nothing, but their absence subtracts. A shortlisted MSP with no recognisable partner status reads as a risk, and buyers are scanning for reasons to cut the list. So you keep the badges for the same reason you tidy the office before a visit. You simply stop expecting them to win anything. The practical consequence is a change of question. Stop asking what else you can add to the credentials strip. Start asking what you could put on the page that a weaker firm would find too expensive to copy. That question has a small number of answers, and every one of them is a commitment rather than a claim.
Section 4
A proof hierarchy, ranked from strongest to weakest
Here is the ranking, strongest first. Argue with the order if you like, but the logic behind it is consistent: how expensive would this be for a bad provider to fake? One. References the buyer picks. You hand over your client list for a segment, or your last ten onboardings, and they choose who to call. A weak provider cannot survive this, because their random client is a coin flip. A curated testimonial costs nothing and proves nothing. Two. A response commitment with a stated consequence. Not rapid response. A defined clock, a defined start, and a credit that appears on the invoice when you miss. Expensive to fake, because faking it means paying every month. Three. A published escalation path with named roles. Who owns the ticket at 30 minutes, who at two hours, who gets called at midnight, and when a client may go over a technician's head without asking permission. A firm with one overloaded senior engineer cannot publish this honestly, because doing so exposes that the answer is the same name three times. Four. A free or capped exit. What it costs to leave, what documentation you hand over, and how long it takes. Providers who rely on switching pain to hold clients cannot copy this, because it removes their retention mechanism. Five. Operating artifacts. A redacted monthly report, a restore test log with real dates, a post-incident review you actually wrote. These take discipline to produce, which is why most firms have none to show. Six. Named clients with specific outcomes those clients will confirm. Weaker than a buyer-picked reference because you chose them, stronger than a badge because they can be contacted. Seven. Certifications and partner badges. The floor. Have them, put them where a buyer expects to find them, and stop there. Most MSP websites are built almost entirely from item seven, and the sales call then does all the work of items one through five verbally. Moving that material onto the page is the actual project.
Section 5
Design the proof so your own self-interest makes it credible
Items two through four in that list work for a reason that has nothing to do with sounding confident. They change your incentives in a way the buyer can verify. That is the design principle: build the proof so that keeping it is cheaper for you than breaking it, and the buyer no longer has to take your word for anything. Take a response commitment. We respond fast costs nothing to say, so it is noise. Compare it with this: a human who names themselves and owns the ticket replies within 30 minutes during business hours, or 5 percent of that month's fee comes off the next invoice automatically, without you having to ask. That is a different object. A firm without a dispatch discipline and an after-hours rota cannot say that sentence without losing money every month. Building one that survives contact with reality takes five decisions. Define the clock start precisely, on something both sides can see, which usually means ticket creation in your system rather than when an email was sent. Define what counts as a response. A named human who owns it. Autoresponders do not count, and saying so out loud is itself a signal, because your competitor's autoresponder is what the buyer has been living with. Cap your total monthly exposure so that one bad month cannot become an existential event. Measure it privately for a quarter before you publish anything. Publish the threshold you already hit 95 percent of the time, not the one you wish you hit. Apply the credit automatically. A credit the client has to chase is not a commitment, it is a complaints process. Then accept the cost honestly. If you never pay a credit, one of two things is true: your threshold is loose enough to be meaningless, or you are not really measuring. Paying out occasionally is the price of the signal being real, and it is usually cheaper than the discount you currently hand over to win the same deal.
Section 6
What to do with the badges you already have
Nothing above argues for deleting your credentials. It argues for demoting them and translating them. Demote them. The logo strip belongs on a credentials page where a buyer who wants to verify you can find it, and in RFP responses where a field asks for it. The top of your service page should carry a commitment, not a wall of vendor marks. Translate them. Every badge that genuinely earns you something should appear as the sentence it buys, not as an image. Compare Microsoft Solutions Partner with this: our partner tier gives us a named escalation contact at Microsoft, so when the outage is on their side your ticket is not sitting in the same public queue as everyone else's. Same credential. The second version is legible to the office manager who is actually making the decision. The same translation applies to firm-level audits. SOC 2 Type II means little to a 40-person client. An outside auditor tested our access controls and change process over 12 months, and we will send you the report under NDA, means something, and it invites a request you can fulfil in one email. This work sits next to your packaging, because the commitments that do the separating have to attach to something the buyer is buying. If your tiers are three columns of feature bullets, there is nowhere to put a response commitment or an escalation path. Building tiers that can carry that weight is covered at https://bizgrowthaxel.com/blog/packaging-managed-services-into-tiers-that-compare-well/, and the wider set of growth patterns for IT service firms sits at https://bizgrowthaxel.com/msp/. The sequence matters more than the design. Decide what you are willing to commit to, then decide where the badges sit. Rearranging logos on a page that promises nothing is redecorating.
Section 7
Where this ages, and what the model cannot see
Every argument like this has a condition under which it stops being true, and this one has a clear one. If cyber insurers or regulators begin mandating specific certifications as a condition of coverage or of serving a sector, certifications stop being marketing and become gating requirements. That inverts the whole argument. The badge still does not differentiate you, because everyone still trading holds it, but not holding it removes you from consideration entirely. At that point the correct response is to get certified immediately, with no reference to marketing value at all. Watch for the early version of this. It shows up as insurance renewal questionnaires asking about the provider's controls rather than only the client's, as prime contractors pushing obligations down to subcontractors, and as sector bodies publishing supplier requirements. When a client forwards you a form you cannot complete, that is not a marketing problem and no proof hierarchy will save you. Treat it as a compliance deadline and staff it accordingly. Now the limit of this entire way of thinking. Signaling and incentive design look at one thing: what an outsider can infer from outside. They are silent on everything a certification does inside your business. They cannot see that a certification track genuinely raises what an engineer can do, and that the exam is often the only forcing function that makes the study actually happen. They cannot see the margin effect of a higher vendor tier across your whole licensing base. They cannot see that good engineers choose employers who pay for certifications, or that a shared syllabus standardises how your team works when nobody is watching. So judge certifications on two separate ledgers. As a marketing asset, most of them are weak. As an operating investment, several of them are strong. A badge can be a poor differentiator and a good decision at the same time. Owners get into trouble mainly by using the second fact to justify the first.
Section 8
Fitness test: are you ready to lead with commitments instead of credentials
You are ready for this if most of the following is true. You can measure response times today from your own system rather than from memory, because you cannot commit to a number you cannot see. You have more than one person who can own an escalation, so a published path does not print the same name three times. You know your gross margin per client well enough to know what a service credit actually costs you in a bad month. You are losing shortlists to firms you believe are worse than you, which suggests the buyer has no way to tell you apart and is deciding on price. And you have a channel where strangers evaluate you before they speak to you. If nobody reads your site before calling, your site is not where the problem is. You are not ready if any of these hold. Your pipeline is almost entirely referral and your close rate is already high. Commitments will still improve your delivery discipline, but scope that as an operations project rather than a marketing one and expect no lift in win rate. Your delivery is inconsistent enough that a published commitment would be a promise you break in month two. A broken guarantee is worse than no guarantee, because it converts a neutral page into evidence against you. You serve mostly compliance-driven or enterprise buyers. For you, certifications sit closer to a gate than to marketing, and the sequence is to satisfy the gate first and worry about differentiation second. Or your binding constraint is capacity rather than demand, in which case sharpening what you promise is the wrong direction. If you land in the not-ready column, the first move is not a website change. Spend one quarter measuring response and resolution times against the commitment you would like to make. You will find out whether you can afford the signal before you publish it, which is the cheapest version of this lesson.