Engage an agency as a teacher

The scope of the page. This is one route that worked, not a judgment that agencies are a bad buy. Large companies use large agencies. The route here is written for a business spending at the lower end and wanting the skill in-house afterwards.

1. What an agency is, and the cycle to avoid

An agency is a lead-generating service business. You pay it to run advertising, do outreach, or package and distribute content on your behalf.

Hired in the ordinary way, the engagement typically runs a cycle:

Stage	What happens
1	They onboard you as a customer
2	Their best person goes on your account, and you get results
3	That person is moved to a new account and a junior takes over
4	Results fall off
5	You ask for the first person back, and they come back part-time
6	They are split between your account and the new one, so results only partly recover
7	You cancel — and then sign up with a different agency, and it starts again

The onboarding at stage one teaches you something, which may or may not be valuable. Everything after it is the cycle. The whole of this page is a way of stopping at stage one deliberately and keeping what you learned.

2. When to use one at all

You could hire an agency to get leads on your behalf. This page does something else with one.

Use an agency when you want to learn a new method or a new platform. That is the trigger, and it maps directly onto the ladder on SOP 40: new placements, then new platforms, then a new method entirely.

The reasons given for why it is faster than doing it yourself:

• The risk of you making a mess of something genuinely new is high.
• They often know basics you do not, and can apply those basics in the shape of your business rather than in the abstract — so you learn faster.
• Because they are an agency, they usually want results out of you quickly, and can get to results faster than you would by trial and error.

And the condition that governs the whole page: this works if you have money. With no money you have to do it yourself, and nothing on this page applies. This procedure is aimed predominantly at small and mid-sized businesses, for whom ten, fifty or a hundred thousand dollars on an agency is a significant cost. Very large companies do use very large agencies; that is a different arrangement and this is not it. And if you have no wish to learn the skill at all, agencies used the ordinary way are fine — this page is for the case where you do want to learn it.

3. Why an agency is a better instructor than it looks

The argument runs like this. Many people who teach a skill are not good at it now, because it has been a long time since they did the work. An agency is in the opposite position: it is doing what works right now, because its business depends on that, and it is usually good enough to have enough clients asking for the work that it genuinely can transfer the skill.

What it lacks is the structure for teaching. That is precisely the gap the offer in §4 fills — you supply the structure and pay for the inconvenience.

4. The offer to make

Say what you want out loud, in this order. The elements below are the ones to include; put them in your own words.

• The want. You would like to do in your own business what they do.
• The admission. You do not know how.
• The term. You would like to work with them for six months in order to learn it.
• The premium, and what it buys. You will pay extra on top for them to break down why they make the decisions they make, and what steps they take to make them.
• What happens next. Once you have a good grasp of how it works, you start training your own team on it.
• The step down. When your team can do it well enough, you move to a cheaper consulting arrangement.
• The reason to give for the step down, in their interest. That way they can still help you when something goes wrong.
• The close. Ask whether they are opposed to doing it that way.

Why it lands. What makes it work is this: you set the expectation at the front, plainly, and you offer to pay more precisely because the arrangement is inconvenient for them. A good number of people will say yes; if one will not, somebody else will.

Negotiating is normal. Expect to negotiate a little to make it fit. If somebody does not want to do it, that is not a problem — ask the next one.

4.1 The working method inside each session

The arrangement is not a demonstration you watch. The terms to set for the sessions themselves:

• They look over your shoulder.
• You are the one driving — your hands on the controls, not theirs.
• Every time something is changed, you want the reasoning: why this change, why now, why they waited before doing it, and why they did not do some other thing.

Record the sessions, take notes, and do whatever work they set you between them.

4.2 The worked case, and the differing durations

The worked case ran differently from the six-month term in §4. Both are set out below.

	The worked case	The term proposed in §4
Who	One solo agency owner	An agency
Why not the agencies first approached	Their fees were, at the time, beyond what could be risked	—
Shape	One hour a week	Not established on this page
Rate	$750 an hour	Their normal fee plus a premium
What was asked for	Some of the more advanced work — retargeting and tracking	A method or a platform
Length	Eight weeks	Six months
Total	$6,000	Not established on this page
Scope	Running ads at a more advanced level	A method or a platform

The arithmetic above holds: seven hundred and fifty dollars an hour across eight weekly sessions is six thousand dollars.

The ask itself was whether the operator could simply show how it was done. The eight weeks split six and two: by week six everything the operator did was understood, and two further weeks were run only to be sure. Then the arrangement ended. The return was one of the highest of any investment made — a judgment reached afterwards, against millions earned with the skill, so it is retrospective and not a forecast. This page gives no figure for the return you should expect.

5. Running it, and the deadline

Prefer the with-me model over the done-for-you model. Ask them to do it with you rather than hands-off. Results come sooner, and by the time their senior person usually rotates off your account you should be good enough to hold those results rather than watch them fall — because you have learned the process and the thinking behind it, applied to your own business.

Set the terms with them, and the deadlines for yourself. The deadline is mostly an internal instrument, pointed at your own team rather than at the agency, because without it you get lazy and dependent. The form it takes is a flat internal commitment: off the engagement by month six.

Expect to pay twice for a while. There will be a stretch where you are paying your own team and the agency at the same time. That is the cost of the method, and the judgment offered — stated as an opinion, not a finding — is that the longer-run cost of not knowing how to run the new platform or method outweighs the short-run cost of paying twice.

6. The exit, in six steps

• Decide whether an agency makes sense for you right now.
• Talk to enough agencies to get a feel for the market, and do not shop on price. How to choose among them is SOP 46.
• Use the agreement in §4.
• Set a clear deadline, to force you and your team to learn the skill.
• Run both teams until your team beats their team regularly.
• Switch to the discounted consulting arrangement, and keep it until you find you are teaching them rather than them teaching you. At that point, end it.

The exit is defined more than once over those steps — a calendar deadline, your team beating theirs, and the direction of teaching reversing.

7. What this page does not decide for you

These are gaps in the procedure, not omissions from this page.

• How much the premium in §4 should be. "Extra" is the instruction; no rate, percentage or multiple is given.
• What the discounted consulting arrangement costs, or how it is structured.
• How to measure "beats their team regularly". No metric, margin or number of periods is given.
• What to do if the six-month deadline arrives and your team is not ready. Both the deadline and the readiness test are stated; what happens when they disagree is not.
• How much money counts as "having means". The condition gates the whole page and no threshold is attached to it.
• What to do if they say no to the arrangement and you have no alternative candidate.

8. The checklist

Question	The answer
Who runs this	You, with your own team
What an agency is	A lead-generating service business
The ordinary cycle	Onboard, senior rep, results, junior rep, results fall, senior rep back part-time, cancel, repeat
The trigger for using one	You want a new method or a new platform
The gating condition	You have money; with none, do it yourself
Who it is aimed at	Small and mid-sized businesses, predominantly
What you are buying	The skill, transferred — not the work, done
The term to propose	Six months
What the premium buys	The reasoning behind each decision, and the steps
Who drives in a session	You do
What you do with each session	Record it, take notes, do the work set between them
The worked case	One hour a week, $750 an hour, eight weeks, $6,000
When the skill was understood	Week six, with two weeks added to confirm
The service model to ask for	With-me, not done-for-you
Who the deadline is really for	Your own team, mostly
What to expect during the handover	Paying your team and the agency at once
The six exit steps	Decide, shop, propose, set the deadline, run both, step down and end
The size of the premium	Not established on this page
The return to expect	This page gives no figure for it

9. What this page does not cover

Choosing which agency to engage — the checks to run on one — is SOP 46. Deciding whether to buy a skill outright in a person instead of renting it is SOP 44, and training your own people once you have the skill is SOP 43. The ladder that tells you a method or platform is new enough to be worth buying help for is SOP 40. Raising a shortlist by word of mouth is the first of the checks on SOP 46.

Contracts, notice periods and fee structures are not covered on this page.