Capture
Google Search.
For categories people already type into Google: the tool, the integration, the alternative to the incumbent. High intent, expensive terms, a negative list built from the job seekers and the students.
Growth
The channel, the offer, the creative, and the measurement that connects a click to a qualified lead, an opportunity, or a first paid action. Four practices, run together, as Growth Partner. Media buying happens inside the system. So does the page. Neither one is the product.
Practice one
Four pieces, in this order. The handoff is part of the system: who replies, and how fast. A channel that produces leads nobody calls is a media buy.
Company type, role, use case, and the disqualifiers. Written in week one, with who you turn away. Every lead after that is read against it.
Search when the category is already typed into Google. LinkedIn or Meta when the buyer has to be introduced. YouTube when the offer needs explaining. One channel until the record says it works. A second is earned at day ninety.
What is promised, to whom, and the one next step: a demo, a trial, a signup, or a purchase. Tested as its own variable, because the same ad with a different ask produces a different buyer.
Message, proof, format. This is the variable that gets tested. Bids and budgets are maintenance. The page the campaign needs ships here: written for the offer, instrumented, changed when an experiment calls for it.
The channel
Chosen by the buyer, not the package.
Capture
For categories people already type into Google: the tool, the integration, the alternative to the incumbent. High intent, expensive terms, a negative list built from the job seekers and the students.
Creation
For a buyer who has to be introduced: by title, company, and stack on LinkedIn, by behavior on Meta. The creative is the test. The page carries the explanation.
Explanation
When the offer needs ninety seconds to understand. Added when the first channel is already producing qualified leads, not before.
Practice two
The source of truth is the CRM or the billing system, not Ads Manager. A result that exists only inside the ad account does not count.
Week one, before spend
Every month
Incrementality is the question “would this have happened anyway?” We name the question and do not sell the test. At the spend most companies start with, a holdout cannot be read. A lift study is a later project, when a channel is large enough to split and still read.
Practice three
One change, one hypothesis, one result, one decision: scale, kill, or iterate. At least one shipped every week. Written down.
The surfaces
Audience, creative, offer, the page, the handoff, and activation if the company can ship a product change. The page is one surface among six.
The rule
“Kill if cost per qualified lead is above this after this many leads.” At a few thousand dollars a month, one change at a time. Sequential tests and a pre-written rule, not simultaneous A/B tests and a significance readout.
The backlog
What was cut, and why, stays in the record. That log is part of the product.
Practice four
Click, a CRM record with the source still on it, an opportunity or a paid event, the outcome written back. When the sales cycle is longer than the term, pipeline is the number, and revenue follows on the company’s cycle. Collected revenue or the first paid action is claimed when it happens inside the term.
Speed to lead is in every report; it is the revenue-connection metric most reports omit. One question in every monthly meeting: why did the last losses not buy? The answer changes the customer definition and the creative. Once the CRM is clean, the ad platform is trained on qualified outcomes, not form fills.
Outside scope
Not included
Why
Each has specialists who do it better than a growth team doing it on the side. We will say when one is the right next move and point you to someone good.
Questions
If yours is not here, ask. We answer plainly, including when the answer is that this is not the right engagement.
The one your buyer is already on. Search when the category is already typed into Google. LinkedIn or Meta when the buyer has to be introduced. YouTube when the offer needs explaining. One channel until the record says it works. A second is earned at day ninety, not added to fill a slide.
Yes, inside the engagement: the message, the proof, the format, cut from what you already have. Creative is the variable that gets tested. Bids and budgets are maintenance. Large productions are scoped separately.
From the click to the CRM. Qualified lead or qualified signup, opportunity, activation, first paid action, each with the source attached. Every month the platform’s conversions sit next to the CRM’s count. We report cost per qualified lead and, once opportunities exist, cost per opportunity, with the month’s fee inside that cost.
At a few thousand dollars a month, one change at a time. The decision rule is written before the test: kill if cost per qualified lead is above this after this many leads. Sequential tests and a pre-written rule. We do not promise simultaneous A/B tests or a significance readout at this spend.
That is the incrementality question, and it is the right one to ask. At the spend most companies start with, a holdout cannot be read, so the answer is the CRM: a result that exists only inside the ad account does not count. A lift study is a later project, when a channel is large enough to split and still read.
Common. Usually the tracking stopped at the form, so nobody could tell which campaign produced the junk. Give us access to the old account. Its history is useful even if its campaigns were not.
The page the campaign needs ships inside the system: written for the offer, instrumented, and changed when an experiment calls for it. A marketing site is a separate conversation, opened only when the record shows the site is the constraint.
No. The fee is flat. A recommendation to spend more has to survive the revenue record, because the invoice does not grow with the budget.
We will say which channel we would start on, and why, on the first call.