Section 1
Your buyer is non-technical and recently burned
Picture the actual visitor. She runs operations at a 40-person accounting firm, is not an engineer, and does not want to become one. She is on your site at 9:40 on a Tuesday because the file server was unreachable for most of Monday, her provider took six hours to answer, and a partner asked her in front of people who chose these guys. Two facts follow, and they set the design brief. The first is that every decision page has to read at a business level. Home, services, pricing, contact. Outcomes, response commitments, and what happens when something breaks, in plain English. Acronyms like EDR (endpoint detection and response, software that watches machines for attacker behaviour) and RMM (remote monitoring and management, the tooling used to patch machines at scale) belong one layer down, where a technical evaluator will look for them. Leading with acronyms loses the person who signs. We have written separately about why that language actively backfires with non-technical buyers at https://bizgrowthaxel.com/blog/why-non-technical-buyers-distrust-technical-language/. The second fact matters more and gets designed for less. She arrived after a failure, so she is not shopping for gain. She is trying to stop a loss from repeating. In that state the risk of the transition weighs heavier than the promised improvement. She can vividly picture a migration week where email breaks and she owns the blame, and she can only abstractly picture your better service. That asymmetry is why capability-first sites stall. Your page can be entirely true and still lose, because the objection it never addresses is not can they do the work. It is what happens to me in the 30 days between providers.
Section 2
Cheap signals separate nobody
Economists call this information asymmetry: one side of a transaction knows something the other side cannot verify. You know whether the work is good. The buyer does not, and cannot find out until months after signing. When that gap exists, buyers stop reading claims and start reading signals. A signal only works if it is expensive or impossible for a weak provider to imitate. Run every element of your homepage through that test. Cheap signals, meaning any competitor can copy them by lunchtime: proactive, enterprise-grade, tailored to your needs, 24/7/365 in a badge, a shield icon, vendor logos, stock photography of server rooms, a team page with no names, certified engineers with no count. Costly signals, meaning a weak provider cannot publish them without getting caught: your average first-response time last quarter and the ticket volume behind it, the number of engineers and where they physically sit, the date of your last tested restore, a written switching plan a prospect can read before signing, a named client who will take a phone call, the exact list of things you will not do. Notice what makes the second list work. Each item costs you something. A published response number has to be measured, republished, and occasionally paid for in credits. A named reference costs relationship capital. An exclusion list loses you the occasional deal on the spot. That cost is the mechanism, not a side effect. A sloppy operation cannot afford any of it, which is why publishing it separates you. The uncomfortable implication: if a claim is comfortable to publish, it is probably doing no work.
Section 3
A homepage rewritten: before and after
Take a 14-person provider in a mid-size metro serving professional firms of 20 to 80 seats, decent retention, losing deals to a cheaper competitor. The before, close to what most sites in this category say: Headline: Enterprise-Grade IT Solutions for the Modern Business. Subhead: Leveraging cutting-edge technology and industry best practices, our certified team delivers proactive managed IT, cloud, and cybersecurity solutions tailored to your unique needs. Call to action: Learn More. Below it, six vendor logos and a padlock. Nothing there is false. Nothing is checkable either. Every competitor in that metro could run the same copy without changing a word, so the page does no sorting. The after: Headline: Managed IT for 20 to 80 seat professional firms in [metro]. We answer in 15 minutes during business hours or the month is free. Subhead: Nine engineers, all based in [metro], supporting 34 client firms. We take over from your current provider on a written 30 day plan, and you can read the plan before you sign anything. Proof line underneath: average first response last quarter, 11 minutes across 1,940 tickets. Backup restores tested monthly, last full test 8 July 2026. Primary call to action: Read the 30 day switching plan. Secondary: Book a 30 minute IT review. Every change does one job. The response number is falsifiable, so it carries weight the badge did not. The seat range and metro tell a wrong-fit buyer to leave, which raises the quality of everyone who stays. The engineer count answers the fear that you are one person with a van. The switching plan attacks transition risk, and offering it as a read rather than a call makes the next step nearly free. Be honest about the cost. Once that number is on the page you own it: measured, republished quarterly, defended after a bad month. A firm that will not do the measuring should not publish it, because a stale or quietly deleted metric reads worse than never having claimed one.
Section 4
Package the invisible: tiers, scope, and pricing signals
Managed services have a presentation problem most services do not. When the work is done well, the client experiences nothing. No outage, no breach, no blown quarter-end. Absence is the product, and absence has no texture, so the buyer compares the only field that does: the monthly rate per seat. There, the cheapest firm wins by definition. Packaging changes what gets compared. Three tiers, named for the buyer's situation rather than for metals, with concrete inclusions: help desk coverage window, response targets by severity, patch cadence, backup and retention period, security monitoring, and a per-seat range. Tiers should differ by how much risk moves from the client to you as the price rises, not by how many bullets each column carries. A deeper treatment sits at https://bizgrowthaxel.com/blog/packaging-managed-services-into-tiers-that-compare-well/. Two things belong on the page that most providers leave off. Scope boundaries, written as three lists rather than one. What is included without limit. What is included up to a stated cap, such as onsite hours per month. What is out of scope and billed separately, such as cabling, hardware procurement, or an application rollout. Publishing exclusions feels like handing ammunition to a competitor. In practice it reads as maturity, because only a firm that knows its cost to serve can draw the line that precisely. Price multipliers, stated plainly. Seat count, device count, number of sites, on-premise servers, after-hours coverage, regulated data, legacy systems in production. The prospect who walks away is usually not the one who found you expensive. It is the one who suspected the quote was an opening figure and the invoices would drift. Different arrivals also need different first screens. A breach scare needs an emergency phone path. Someone leaving an unresponsive provider needs the switching plan. Someone outgrowing a single internal IT person needs the tier comparison and evidence of team depth. Someone under regulator pressure needs your compliance scope. Give each a named entry point above the fold.
Section 5
Security credibility: the proof hierarchy for IT providers
Security is now a common reason businesses go looking for a provider, and it is where empty claiming is thickest. Every MSP site says proactive security. Rank your evidence by how hard it is to fake, and put the hard end where decisions happen. Weakest, close to worthless: adjectives and icons. Military-grade, bank-level, next-generation. These cost nothing to write. Weak: vendor partner logos and individual certifications. Real, but widely held. They say you cleared a bar thousands of firms also cleared. Keep them, place them low. Stronger: audited attestations tied to a date and a scope. SOC 2 Type II or ISO 27001, stated with the audit period and what the scope actually covered, report available under NDA. The date and scope carry the signal. A logo with neither drops back into the previous category. Never present a framework you are working toward as one you hold, because an insurer or a client security questionnaire will eventually check. Strongest: things only a competent firm can put in public. A written incident narrative with real dates and numbers, such as a 30-person firm hit by ransomware on a Friday, contained within a stated number of hours, restored from a backup tested days earlier. A named client in the buyer's industry who will take a call. Your own internal practice published as policy: MFA on every administrative account, restores tested monthly with the last test date shown, a summarised incident response runbook. That last category is expensive. Incident stories need client permission and usually a legal review. Most firms will not do it, which is the entire point of the signal. One more thing quietly undermines all of it. A slow page, a broken form, a dead link, a copyright line reading 2023, a subdomain throwing a certificate warning: each one tells the buyer something about your maintenance discipline that no testimonial reverses.
Section 6
Your website is your first SLA
An SLA, in plain terms, is a written promise about how fast you respond, with a consequence attached when you miss. Your site is where the buyer forms an expectation about that promise, and it is the first live sample of your operational discipline they observe. Consider what happens. A prospect fills in your contact form at 4:52pm on a Friday, having just read a headline promising 15 minute response. Your reply lands at 10:14am Monday. Nothing was technically violated, since the business-hours SLA was never about web forms. But the buyer ran an unintentional test and got an answer that contradicted the headline. Every other claim on the site is now discounted. So wire the intake to the promise. An automatic acknowledgement that names a real person and states a time window. A visible phone number for anything urgent, because a business in an outage will not wait for email. A booking link so a qualified prospect can take time without a round trip. And a plain statement of what happens outside business hours, even when the honest answer is next business morning. Stating a limit costs less than implying a capability you do not staff. Then hold the site itself to the standard you sell. Page load, uptime, TLS configuration, DNS hygiene, form deliverability, a monitoring alert on your own domain. Give one person accountability for it, the way you would for a client tenant. If you monitor your own site with the tooling you sell, say so. That is a checkable detail a weak competitor will not copy, because they are not doing it. This is where web design stops being a design conversation and becomes an operations one. The site can only make promises your delivery can keep, so the sequence runs backwards from how most projects get scoped: fix the operational commitment first, then publish it.
Section 7
When the audience inverts, and what none of this sees
Everything above assumes a non-technical buyer. There is a clean break in that assumption, and on the wrong side of it the advice reverses. If the buyer has in-house technical staff, an IT manager, a systems administrator, a CTO, or a compliance officer running a vendor assessment, the audience inverts. Plain-English simplification now reads as thin. That person is defending their own judgement to their leadership, and their evidence is depth: your RMM and PSA stack by name, documentation standards, change management, patch testing, escalation matrix, a sample monthly report, security controls mapped to whichever framework they answer to. A page written for the non-technical owner gives them nothing to carry into that meeting. The rule: build both layers, and let your pipeline decide which is primary. If co-managed work, where you operate alongside an internal team, is a meaningful share of what you win, the depth layer is your main page. If it is a small share, keep depth one click down and let the business-level pages lead. Now the blind spot, and it is a large one. None of this sees referral-driven deals. If most of your work arrives through an accountant, a vendor rep, or a client telling a peer, the decision was substantially made in a conversation you were not in. The site is then a credibility check. It does not have to convert. It has to avoid creating doubt: current, fast, coherent, real names and faces, and confirmation of whatever the referrer said about you. Building elaborate conversion architecture when four in five deals are referred wastes money and attention. Find out before you spend. Ask every closed deal how they first heard of you and what they looked at before the call, then let the answers set the budget.
Section 8
Fitness test: is your site the right thing to fix
You are ready to rebuild along these lines if most of the following is true. You can measure at least one operational number and stand behind it publicly, such as first response time or restore test frequency. You have a written switching plan you run rather than improvise. You can name a segment tightly enough to put it in a headline, by seat range, industry, or geography. You can staff every commitment you would publish. You have cold traffic, meaning strangers evaluate you before speaking to you, because that is where signals do their work. And you are losing deals on price to firms you believe are worse, which points at the comparison frame rather than the price. You are not ready if any of these hold. Your pipeline is nearly all referral and it is full, in which case fix intake speed and hygiene and spend the rest elsewhere. You cannot describe your ideal client without listing four industries. Your delivery cannot support a published response window this month, so fix delivery first, because publishing a commitment you then miss is the one change that reliably makes things worse. You are mid-acquisition or changing your service model, so anything you publish now will be wrong in a quarter. Or you want the site to resolve a positioning problem, which it cannot, since a website only expresses a decision already made. A closing caution about advice in this category, including ours. Most website guidance for MSPs is written by people who sell websites, and that shapes what gets recommended. We sell them too. Our MSP page at https://bizgrowthaxel.com/msp/ is built on the framing above and is worth reading as a worked example. The test that matters is still the one at the top: for every element you are about to pay for, ask whether a buyer could check it, and whether a weaker competitor could claim it for free.