When PMs talk about acquisition, the conversation often collapses far too quickly into a few familiar questions: what is CAC, what is ROAS, which channel converts better?

Those questions are not wrong. They are simply downstream of the real problem.

Acquisition is not one thing. It is a bundle of very different jobs jammed into the same spreadsheet. SEO deals with discoverable demand. Paid often acts as a fast learning loop for message, audience, and landing page fit. Partnerships can borrow distribution, context, or trust. Attribution allocates credit, but that is not the same as causality. Incrementality is closer to the harder question: if we had not done this, how much would the result actually have changed?

So this is not a grand tour of every acquisition tactic under the sun.
What I want to do instead is separate a set of concepts that growth teams routinely mix together, then show where each one is genuinely useful and where it starts to mislead.

My argument, up front

Acquisition gets easier to manage once you stop treating every channel as the same funnel input and start asking what each one is actually for, then decide whether attribution is enough or whether you need incrementality.

That sounds obvious when written down. In practice, teams still trip over it all the time.

If you treat SEO, paid, and partnerships as nothing more than competing rows in a channel table, you tend to create three kinds of confusion:

  1. brand demand grows and paid claims the final click;
  2. the product gets better and several channels benefit at once;
  3. one channel is excellent at closing demand that already exists, but weak at creating new demand.

That is how teams end up working very hard, seeing tidy dashboards, and still failing to answer the only question that matters: what actually created more customers?

First, split the stack into five layers

I find it useful to separate acquisition into five layers.

1. SEO: making existing demand easier to find

SEO is best when the demand already exists, can be searched for, and can compound over time.
It is less like a campaign and more like a durable information architecture.

2. Paid: buying speed to test message, audience, and distribution

Paid is valuable not merely because it can buy traffic, but because it can shorten the learning loop.
You can test message, audience, and landing promise faster than you usually can through product changes alone.

3. Partnerships: borrowing someone else’s distribution, trust, or surface area

Partnerships are not just “doing a collab”.
The useful ones typically exchange one of three things: distribution, product context, or trust.

4. Attribution: allocating credit across an existing path

Attribution is an accounting system.
It helps you assign credit, but it does not automatically answer whether a channel caused new conversions to happen.

5. Incrementality: checking whether a channel generated additional outcomes

Incrementality is much closer to a causal test.
It is trying to answer not “who gets the credit?” but “what would have happened without this channel?”

Once those layers are mixed together, the discussion becomes mush.

SEO is not merely content production. It is demand capture through findability

Google’s own documentation is refreshingly plain on this point. Search is a fully automated system, and Google does not accept payment to crawl a site more frequently or rank it higher. That matters because it reminds us that SEO is not a slower, weirder version of media buying. It is the work of making content easier to crawl, index, and understand.

That means a PM should not start SEO discussions with “how many sessions will this article bring?”
A better starting point is three sharper questions.

First, does the page map to a real user task?

In a bookings product, “late check-in hotels near the airport” or “family-friendly hotels without the usual traps” are more useful entry points than another page about how brilliant your platform is.

Secondly, does it sit inside a sensible information architecture?

Google repeatedly stresses crawlable links, link architecture, and understandable URL structure.
What many marketers call a topic cluster is useful, not because the phrase itself is magical, but because it forces you to build a coherent parent-child structure with internal links that both users and crawlers can follow.

Thirdly, does the content lead back to product value?

This is where many content programmes quietly stall. Traffic is not terrible, publication is regular, but the pages float in a separate universe from the product.
For a PM, SEO should not mean “publish more”, but rather “build content paths that connect discovery to product value”.

When SEO deserves to be a priority

  • the demand can already be searched for;
  • users do research before deciding;
  • your value proposition can be understood without a sales call;
  • you can accept a slower but compounding payoff.

When not to make SEO the main battlefield

  • the category still needs to be taught to the market;
  • users do not yet know what to search for;
  • the team cannot maintain content properly;
  • you need feedback inside two weeks rather than over a quarter.

The problem is rarely that SEO does not work. It is that teams use it as a short-term fire extinguisher.

Paid is often reduced to “throw budget at traffic”.
For PMs, a more useful way to see it is as a fast testing environment.

If you already have several candidate messages, audience cuts, or landing promises, you do not necessarily need to hard-code them all into the product first. A small paid test can tell you which direction deserves further investment.

The important bit is not how much money you spend. It is whether you control the variables.

A minimal paid test should lock down a few things first

  • change one primary variable only: message, audience, creative, or landing page;
  • define success before launch, not after the chart moves;
  • keep tracking clean: UTMs, event definitions, and platform tags should not be fighting each other;
  • give the test a sensible window instead of declaring victory after half a day.

Google Ads documentation contains a small but very practical note here: in GA4, auto-tagging takes priority over manual tagging. That sounds like plumbing, but it matters. Teams often argue about channel performance while the underlying tracking is already muddled enough for Google Ads clicks and GA4 reporting to drift apart.

What paid is especially good at testing

  1. Message
    Which promise lands: saving time, reducing uncertainty, helping people avoid bad choices?

  2. Audience
    Is the offer resonating with business travellers, families, or budget-conscious solo travellers?
    Not every conversion problem starts inside the product.

  3. Landing experience
    Should users see a checklist, an example result, or the search interface immediately?

When not to scale paid yet

  • you still do not know which message has real pull;
  • your tracking is messy;
  • activation is weak, so you are paying to fill a leaky product;
  • the channel appears strong mainly because it harvests demand that already exists.

That is when spend can rise faster than learning quality.

Partnerships are not a side channel. They often move trust and context

Partnerships are easy to describe badly.
People reach for phrases like “strategic collaboration” or “co-marketing opportunity”, and the discussion drifts into fluffy territory almost immediately.

I find it more useful to split partnerships into three buckets.

1. Distribution partnerships

A partner helps you reach an audience you would struggle to reach directly.
Think travel publishers, loyalty portals, credit card travel hubs, or employee benefit programmes.

2. Integration partnerships

The point is not to send more traffic, but to place your product inside an existing workflow.
Calendar integrations, maps, travel management tools, and comparison layers often sit here.

3. Trust partnerships

The audience does not know you well enough yet, but they already trust the other side of the relationship.
This matters especially in low-frequency, high-risk, or high-value decisions.

A partnership becomes interesting when it changes a particular part of the funnel in a structural way. One deal can improve qualified traffic, another can increase activation, another can reduce hesitation before payment.

The common partnership failure mode

  • measuring exposure instead of downstream movement;
  • agreeing on format but not on shared value;
  • launching without dedicated landing pages, offers, or tracking;
  • calling it a partnership and then measuring it as generic direct traffic.

If it cannot be measured, it quickly turns into a pleasant anecdote rather than a repeatable lever.

Attribution is an accounting system, not a causal one

GA4 defines attribution as assigning credit for important actions to ads, clicks, and other touchpoints along the user’s path. That is a useful definition precisely because it is narrower than many teams pretend.

Its value is obvious. Without attribution, cross-channel reporting becomes chaos.
Every channel claims it “influenced” the result, and nobody can reconcile the numbers.

Its limitation matters just as much.
Attribution allocates credit inside an observed path. It does not prove that a channel caused additional conversions to exist.

That is why last-click reporting can bend decision-making so badly.

When attribution is useful

  • you want to understand the touchpoints appearing in conversion paths;
  • you want to see which channels tend to assist earlier versus close later;
  • you need a common reporting language across teams.

When not to treat attribution as the final truth

  • brand demand is already strong;
  • channels heavily overlap;
  • the question is budget allocation rather than bookkeeping;
  • you are trying to distinguish demand creation from demand capture.

For PMs, attribution is most useful as a shared ledger and a prompt for where deeper causal checking may be needed.

Incrementality is the question you ask when you care about causality

Google Ads Conversion Lift and Meta’s Conversion Lift both aim at the same class of problem. They compare treatment and control to estimate how many conversions were truly added by exposure to ads.

That is the crucial distinction from attribution.
Incrementality is asking about additional outcomes.

A very common misunderstanding

A channel with excellent ROAS does not necessarily create much new demand.
Sometimes it simply catches people who were already on their way to converting.

Brand search often behaves like this. In an attribution report it can look brilliant. Once you switch to an incrementality lens, the truly additional impact may be smaller than the report implied.

When incrementality should move up the agenda

  • you are deciding whether to increase channel spend;
  • you suspect a channel is cannibalising organic or branded demand;
  • you are comparing the real value of SEO, paid, and partnerships;
  • the question is new customer acquisition rather than tidy reporting.

When not to reach for incrementality first

  • tracking is still unstable;
  • volume is too small to read clearly;
  • you have not even identified where the product funnel breaks;
  • the team needs message and audience learning more urgently than media calibration.

Incrementality is not something every company needs every week. But once the question shifts from “which touchpoint closed the sale?” to “which channel created additional sales?”, you need a different tool.

A more practical PM decision framework

Suppose you run a bookings product. I would frame the choices like this.

If the problem is “we do not get enough high-intent traffic”

Start with SEO.
You may simply not be capturing search demand around real tasks such as airport stays, late check-in, or family travel constraints.

If the problem is “we do not know which message pulls”

Start with paid.
It is the fastest environment for testing promise, audience, and landing structure.

If the problem is “the trust barrier is too high”

Start with partnerships.
You may not need more reach. You may need borrowed credibility.

If the problem is “every channel claims success”

Start with attribution.
The team needs a common ledger.

If the problem is “should we keep increasing spend?”

Start with incrementality.
That is now a causal question, not a bookkeeping one.

Once you sort the problem that way, you are much less likely to use the wrong tool for the wrong question.

What this piece is deliberately not trying to do

This article deliberately does not go deep on:

  • platform-specific media buying tactics;
  • a full SEO production playbook;
  • MMM or MTA in depth;
  • bid strategy mechanics;
  • complicated partnership org design.

Its job is not to become an acquisition encyclopaedia.
Its job is to separate concepts that are too often flattened into one dashboard.

Closing thought

A lot of PMs start by treating acquisition as a longer conversion funnel.
After a bit of real work, it becomes clear that it is closer to a map:

  • SEO captures discoverable demand;
  • paid accelerates message and audience learning;
  • partnerships borrow distribution, context, or trust;
  • attribution allocates credit;
  • incrementality checks whether the outcome was truly additional.

Once those layers are separated, you can ask a much more mature question:

Am I trying to improve demand capture, distribution efficiency, credit allocation, or real incremental growth?

The next piece takes that question one level higher.
Because mature PMs do not just keep generating ideas. They turn metrics, experiments, decisions, and scale into an operating rhythm that can survive beyond any single campaign.