I took on a managed IT provider earlier this year. Multiple offices, mid-market clients, the kind of shop that does fifty-plus seats and real compliance work. Before I changed anything I did what I do with every new account, which is pull twelve months of Search Console and sort every click by what the person actually typed.
Ninety-two percent of their search clicks were brand, careers, and support.
People typing the company name. People looking for the jobs page. Existing clients trying to find the support portal. Every one of those clicks is a person who already knows the company exists. None of them is a buyer discovering an MSP for the first time.
That left eight percent. Eight percent of a year’s search traffic was doing the entire job of new business.
If you run an MSP and you have never split your traffic this way, run it this week. I have yet to see a managed IT provider whose number comes back healthy.
Why the MSP traffic report looks fine when the pipeline is empty
Here is what makes this hard to spot. Every number on the dashboard moves in the right direction.
That same client’s first month with us: clicks up 12% while impressions fell 17%, click-through rate up from 0.58% to 0.78%, sixty tracked key events including fourteen meetings scheduled. Real movement. Genuinely encouraging.
But if 92% of the base is brand and support, a 12% lift is mostly more of the same people finding the same portal faster. The dashboard rewards you for improving the traffic you already own. It says nothing about whether anyone new is showing up.
MSP marketing has a specific version of this problem that other B2B categories do not. Your existing clients search for you constantly. They search for the helpdesk. They search for the phone number. They search for the remote support download. A hundred-seat client generates search volume all year, and every bit of it lands in the same report as the prospect who typed “managed IT provider” for the first time.
So the metric that matters is the share of clicks from people who did not already know your name. Total clicks will lie to you all year.
Splitting the number takes about twenty minutes
In Search Console, pull twelve months of query data. Pull it at 25,000 rows, because the 500-row default truncates the tail and the tail is where discovery lives. Then bucket every query:
Brand. Your company name and every misspelling of it. Also your name plus a word: “acme it support”, “acme login”, “acme careers”.
Support and existing-client. Portal, ticket, remote, download, phone number, password reset. Every one of these is an existing customer looking for the helpdesk.
Careers. Jobs, hiring, careers, plus your name and a role title.
Discovery. Everything left. “managed it services chicago”. “msp for law firms”. “co-managed it”. “soc 2 compliance it provider”. Someone with a problem, no vendor in mind.
Add up bucket four. That percentage is your actual top-of-funnel. Everything else is your existing book of business generating search traffic that a marketing report will happily count as growth.
What a healthy 8% actually produces
I want to be careful here, because the small number can read as bad news and it usually is not.
A medical courier client of mine runs 30 metros across 8 states. Only 22% of their search clicks are new discovery. That is roughly 389 discovery clicks a month against a much larger total.
One of those clicks, from a city page in Jacksonville, turned into a client now approaching $3,000 a week in accounts receivable. Roughly $150,000 a year from a single search. That client then pulled them into Fort Myers and Madison, Wisconsin, two markets that already had pages built and ranking before anyone needed them.
Three hundred and eighty-nine discovery clicks a month sounds thin. One of them was worth $150,000.
That is the shape of B2B search, and it is why volume is the wrong lens for an MSP. The whole job is being findable at the moment a 60-seat manufacturer decides their current provider has missed one SLO too many. Ten thousand extra visitors would do nothing for that. That happens a handful of times a month in your territory. You need to be there for those.
Build pages around the buying question
Most MSP sites are organized around what the company sells. Managed services. Cybersecurity. Cloud. Backup and disaster recovery. VoIP. Every competitor has the same five pages with the same five stock photos.
Nobody in the discovery bucket types “managed services”. They type the problem they have, the comparison they are weighing, or the requirement they were just handed.
Three page types earn discovery clicks in this category:
Comparison pages. “X vs Y” for the providers you lose deals to, and for the in-house-versus-outsourced decision. This is the single most productive tactic I run anywhere. A national property tax appeal firm I work with went from 5,168 clicks in June to 7,836 in July, up 52%, on the back of comparison and county-level pages. Their click-through rate went from 0.92% to 1.20% and their average position improved from 8.2 to 7.8 in the same month. Traffic normally dilutes both of those when it grows. Concentrating them at the same time is what happens when the page matches the question exactly.
Requirement pages. Compliance is the MSP version of a buying trigger. A prospect who has just been handed a client questionnaire about their security posture searches for the requirement by name, not for “IT support”. Write to the requirement, describe the process honestly, and skip the guarantees. Nobody credible promises 100% on a compliance outcome.
Vertical pages. “IT support for manufacturing”. “MSP for medical practices”. Volume per page is low. Intent is close to perfect, and the search terms carry real commercial weight. msp seo alone runs an $89.43 cost per click, which tells you what the people bidding on it think a lead is worth.
The keyword numbers back this up. msp marketing gets 720 searches a month at a difficulty of 2, and the average page ranking in the top ten has 14.8 referring domains. Fourteen. That is a category where good pages still win on merit.
The shortlist is now built inside an AI answer
The other thing that changed this year: a meaningful share of vendor shortlists are assembled by an AI assistant before a human ever opens Google.
The property tax firm I mentioned is cited by ChatGPT as one of two sources on “best property tax appeal service”, and logged 146 measured AI referral sessions in July. My medical courier client holds a 93.9% presence rate in Google AI Mode on “medical courier” at an average rank of 1.39, and 87.8% in ChatGPT on “stat medical courier”.
Those numbers answer one question. Does the model say your name when a buyer asks who to call. For an MSP, the question sounds like “who are the best managed IT providers in Tampa for a 75-person company that needs SOC 2 support”. Somebody in your territory typed a version of that this week. Either your name appeared or it did not.
What I would do in the first ninety days
- Split the twelve-month query data into the four buckets. Write down the discovery percentage. That is your baseline.
- Check whether you appear in ChatGPT, Perplexity, and Google AI Overviews for the three questions a buyer would actually ask. Record the answer, because you cannot improve a number you never wrote down.
- Ship three comparison pages against the providers you lose to most.
- Ship one requirement page for the compliance framework your best clients get asked about.
- Re-run the bucket split at day ninety. The total-clicks number may barely move. The discovery share is the one that tells you whether the work landed.
I sold my last company, an e-learning business in the construction trades, to private equity in an eight-figure exit. We had 360 keywords ranked in the top three across 156 product pages by the end. The search asset was a material part of what the buyer was paying for. I was the owner carrying the risk on that one, and it taught me something an agency seat never would. The traffic number on the dashboard and the traffic that changes a valuation are two different columns.
For a deeper look at how this fits together, start with our AI SEO agency overview, or read how we approach B2B SEO more broadly.
Find out what the models say about you
The fastest way to see the gap is to look at what an AI assistant says when a buyer asks for a recommendation in your market.
That is what our free AI visibility scan does. We run the questions your prospects are actually asking across ChatGPT, Perplexity, and Google’s AI results, and send you back what came out, including which competitors got named instead of you.
No pitch attached. Run it, look at the answer, and decide from there.