Romario Martin - B2B Agency

Romario Martin - B2B Agency BAD 2 BADASS | The Google Ads Agency That's Accountable to Revenue

Helping Remodeling Companies Turn Google Ads into Revenue & Customers. Need More Customers?

Let's Talk. 🛠️

$68m+ in Spend Managed | $550m+ in Attributable Revenue

08/14/2026

Every agency pitch I have sat through had a senior person in the room who was never seen again.

I know how that works because I was on the other side of it for years. The person who wins the account is expensive and does not scale, so the account goes to whoever has capacity. The senior name stays on the org chart and shows up for the quarterly.

Nobody is being dishonest about it. The arithmetic just runs one way. An agency that grows headcount has to fill that headcount with people who cost less than the work is billed at, and that math resolves the same way every time.

So when I went out on my own, the thing I was actually trying to keep was the part that gets handed off first.

Three things stay with me on every account.

→ The audit. Reading someone's search terms and campaign structure is where you learn what the account is really doing, and it does not transfer through a document. Handing this off means the person making decisions has never seen the raw material.

→ The rebuild. The structure decisions made in the first three weeks determine what is possible for the next year. This is the highest-leverage work in the engagement and it is the work most often given to the newest person.

→ The conversation where I say something is not working. That one is uncomfortable, which is exactly why it gets delegated. It should not be.

The obvious objection is fair and I will make it before you do. One person doing all of that is a single point of failure, and in cybersecurity of all categories, a single point of failure is a thing you would never ship.

Which is why the other half matters. Everything I do gets written into a playbook the client owns and a dashboard that runs without me. Not so someone else can do the work. So that if I get hit by a bus, the account does not restart from zero and the next person inherits reasoning rather than a login.

The senior person does the work. The system means the work is not trapped in their head.

Those two things sound like they are in tension. Building it so they are not is most of what I have spent the last year on.

08/13/2026

The most common thing founders tell me about Performance Max:

We turned it on for efficiency. Now most of the leads are junk and nobody knows how to unwind it.

Performance Max is doing exactly what it was told. That is the problem.

If your conversion action is a form submission, you asked Google to find people who submit forms, as cheaply as possible. It has more inventory than any other campaign type, so it got very good at that very quickly.

The people who fill in forms most readily are students, researchers, job seekers and consultants. A buyer with budget is more careful with their contact details, not less.

The campaign succeeded. The instruction was wrong.

Which is why the unwind usually is not "turn it off." It is changing what the account counts as a conversion, then deciding how much budget Performance Max deserves once it is chasing the right thing.

An automated campaign will always find you more of whatever you told it to value.

08/12/2026

Ross Haleliuk: "The biggest challenge of cybersecurity startups is not the product; it's distribution."

I see the moment that becomes true from an unusual angle. I do not build security products. I spend my days inside the ad accounts of companies that do.

It always looks the same. The first ten customers came from the network. That felt like product validation. It was partly network validation, and the network runs out.

Then someone buys distribution for the first time, discovers the category keyword costs $50 to $200 a click, and concludes three months later that paid search does not work in security.

What got tested was the most expensive way into the channel, run by someone doing it alongside four other jobs.

Nobody would ship a detection engine after a weekend of research. Plenty of companies enter a $100-a-click auction on about that.

Worth separating two questions before you write the channel off: whether it is wrong for your category, or whether it was built by someone who had done it in your category before.

08/11/2026

I am giving away the negative keyword list I apply on every cybersecurity account I rebuild.

8 categories, roughly 250 terms, match type noted on each section.

It is the same list that goes into paid engagements. There is no gated-down version of it.

The reason I am comfortable doing that is that the list is the easy part. There are already published negative keyword lists with thousands of entries, and pasting one into your account is one of the faster ways to damage it.

Here is what those lists are missing, and it is the section I would read first.

A do-not-negate list.

Somewhere in every big scraped list are terms that are buyer signals in this category. I have seen `pricing`, `cost`, `demo`, `enterprise`, `platform` and `vendor` all sitting in lists people paste in wholesale. Every one of those is a qualifying word that tells you a business is doing the searching.

The most expensive one is `vs`.

A lot of informational junk contains it, so someone notices the pattern and negates it account-wide. That single line blocks every search of the form "[competitor] vs [your product]," which is close to the highest-intent query anyone will ever type about your category. Somebody comparing two named vendors is at the end of a process, not the beginning.

I have opened accounts where that one negative had been quietly running for a year.

So the list I am sending has both halves. What to block, and what to protect. The protect list is shorter and it matters more.

Comment NEGATIVES and I will send it over.

Link in the comment section 👇

I have seen paid search accounts get objectively better on paper while becoming worse for the business.Cost per lead dro...
08/10/2026

I have seen paid search accounts get objectively better on paper while becoming worse for the business.

Cost per lead drops.

Lead volume goes up.

The dashboard turns green.

And then somebody from sales asks the question that ruins the meeting:

"How many of these actually became customers?"

That is usually where cost per lead stops being useful on its own.

Take two campaigns.

One generates leads for $100.

The other generates them for $500.

If you stop there, the first campaign is five times more efficient.

But keep following the leads.

The $100 campaign produces 100 leads, five qualified opportunities and one customer.

The $500 campaign produces 20 leads, eight qualified opportunities and three customers.

You paid considerably more for every lead in the second campaign.

You also acquired customers far more efficiently.

That is the part most paid search reporting misses.

The ad platform is very good at telling you what happened before the form was submitted.

→ How much the click cost
→ Which keyword generated it
→ What campaign produced the conversion
→ How cheaply you acquired the lead

The interesting part happens afterwards.

Did sales accept it?

Did it become an opportunity?

How much pipeline did it create?

Did it close?

What did acquiring that customer actually cost?

Those answers normally live inside the CRM, which is why two teams can look at the same paid search program and reach completely different conclusions.

Marketing sees the $100 lead.

Sales sees the person who never replied.

Both are technically correct.

They are just measuring different points in the journey.

This becomes especially important in cybersecurity and enterprise software, where one customer can be worth considerably more than hundreds of form submissions.

A $700 lead that becomes a six-figure customer is not expensive.

A $70 lead that never becomes pipeline is.

After managing $68M+ in paid search spend, this is one of the distinctions I care about most.

I am not trying to make leads cheaper.

I am trying to make customer acquisition more efficient.

Sometimes that lowers CPL.

Sometimes it increases it.

I am fine with either.

08/10/2026

I have managed $68M in ad spend and I still cannot tell you what a good cost per lead is…

Not without knowing what happens after the form.

Someone asked me this last week. Series A, security product, wanted a number to take to a board meeting. What should we be paying per lead?

The honest answer is that the number is unknowable on its own.

Picture two cybersecurity companies. Same category, same monthly spend, same $180 cost per lead.

The first one gets leads from broad category terms. Most are students researching a certification, some are vendors prospecting, a few are real buyers. Sales stops picking up by week three.

The second one gets leads from comparison and alternative terms. Fewer of them, and they take longer to convert on the first touch. Sales books meetings off nearly all of them.

Same $180. One of those is a good number and one is a slow leak, and cost per lead cannot tell you which one you have.

The number that can is cost per qualified opportunity. Getting it means wiring your CRM back into the ad account so the platform can see what happened after the form or call. Most Series A accounts I look at have never done this, which means their bidding has been optimizing toward whoever calls or fills in a form most easily.

Until that connection exists, cost per lead measures how cheaply you can generate activity.

I have watched a lot of accounts hit their cost per lead target and miss their pipeline number in the same quarter.

Search "cybersecurity Google Ads CPC" and you get one number. $35.80.It comes from a single agency page and everyone quo...
08/07/2026

Search "cybersecurity Google Ads CPC" and you get one number. $35.80.

It comes from a single agency page and everyone quotes it, including people building budgets against it.

I went looking for where that number came from, because it did not match anything I have seen running paid search in this category for the last decade.

It is an average. That is the problem with it.

The categories inside security do not price alike. A click in zero trust runs about $20 to $33. A click in pe*******on testing runs $123 to $189. Those are different purchases in different auctions with different competitors, and the gap is wide enough that one average describes neither.

Three things collapse into that $35.80:

→ Brand and non-brand blended together. Bidding on your own name costs a few dollars. Bidding on your category costs many times that. Any account average is really a measure of how much brand traffic that account has.

→ Categories averaged against each other. Pentest, compliance, AppSec, cloud security, SIEM and identity all sit in separate auctions with separate competitor sets.

→ Career and student traffic sitting in the same keyword space. This one is specific to security. Search demand here is dominated by people who want a job in it, not people who want to buy it. Those clicks are cheap, they land in accounts constantly, and they drag the average down while producing nothing.

That last one is worth sitting with. A low average CPC in cybersecurity is often the signature of a waste problem, not an efficiency win.

So I published the distribution instead of the average. 15 categories, real ranges, and the reason each one prices the way it does.

Comment BENCHMARKS and I will send it over.

Link in the comment section 👇

A founder told me his category was too expensive to advertise in.He had priced the obvious keyword. The one with his pro...
08/06/2026

A founder told me his category was too expensive to advertise in.

He had priced the obvious keyword. The one with his product category in it, the one every competitor bids on, the one that costs somewhere north of $50 a click and runs as high as $189 in the most expensive corners of security.

At a $6,000 monthly budget, that keyword buys you roughly 100 clicks. If 2% of them convert, that is 2 leads, or $3,000 a lead. If half of those are real, you are paying $6,000 for one real conversation.

He was right that the math does not work. He was wrong about which math to run.

Category terms are where the whole market bids, so the price reflects the whole market's willingness to pay, including the Series D company with a $400,000 quarterly budget and a 9 month payback tolerance. You are not competing on relevance there. You are competing on balance sheet.

The queries next to it price very differently.

→ Comparison terms. Someone typing "[competitor] vs" has already accepted the category and is choosing a vendor. Often $8 to $25.
→ Alternative terms. "[incumbent] alternative" is a person actively looking to leave. Same range.
→ Problem terms. The symptom they would describe before they knew a category existed.
→ Compliance and requirement terms. "SOC 2 evidence collection" is a person with a deadline.

Lower cost, and closer to a decision. The category term catches people at the beginning of a process. The comparison term catches them near the end of one.

Most early stage accounts I open have spent the majority of their budget on the first group and barely touched the second.

The interesting part is that the second group is usually small. A few hundred searches a month, not a few thousand. Which is exactly why the enterprise players ignore it, and exactly why it is available to you.

08/05/2026

"I know some of this budget is being set on fire. I just cannot see which part."

A head of demand gen said that to me and I have thought about it every week since.

She was not worried about the money. She had raised, she had budget, she was allowed to spend it.

What bothered her was standing in front of a board defending a number she could not fully account for.

That feeling is usually accurate. In the accounts I look at, 40% to 60% of search spend goes to queries that were never going to become a customer.

None of it shows up as a problem. Those clicks have a fine click-through rate. Some of them fill in a form, because a curious person will trade an email for a whitepaper.

So the report looks healthy while the pipeline stays thin.

If you have that feeling about your own account, it is worth finding out whether you are scaling a machine or a leak before you approve the next increase.

Address

Madison, NJ
07940

Opening Hours

Monday 10am - 5pm
Tuesday 10am - 5pm
Wednesday 10am - 5pm
Thursday 10am - 5pm
Friday 10am - 5pm

Website

Alerts

Be the first to know and let us send you an email when Romario Martin - B2B Agency posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to Romario Martin - B2B Agency:

Shortcuts

Share

Category