Understanding adstock: How marketing impact unfolds over time

One of the most common questions we hear from marketing teams is:
“If I spend money this week, when will I actually see the results?”

Marketing rarely behaves like a light switch. Some tactics drive impact immediately. Others build momentum that compounds over weeks. And some—especially brand marketing—create a “halo” that may last months.

This is exactly the concept adstock helps us measure.

In this article, we’ll break down:

• What adstock is

• Why it matters in Marketing Mix Modeling (MMM)

• How to think about decay (how long marketing effects last)

• Why there is no universal answer—and that’s okay

• How Paramark approaches adstock scientifically and pragmatically

What is adstock?

Adstock captures the idea that marketing has a cumulative and lagged effect.

When you spend in a marketing channel, the impact doesn’t happen only on the day or week you spent the money. Some portion of the effect carries over into future weeks, and that carryover decays gradually.

Think of adstock like dropping a pebble into a pond:

• The moment the pebble hits the water is the immediate impact.

• The ripples that continue spreading afterward are the lagged effects.

In an MMM, these “ripples” are mathematically represented using an adstock function that spreads each week’s impact into the weeks that follow.

Why adstock matters in MMMs

If we ignored adstock, an MMM would assume:

• A TV or Meta impression only matters the moment it’s served

• A prospect sees an ad, decides immediately, and that’s it

• Multi-week campaigns have no cumulative effect

All of which we know is wrong in practice.

Adstock helps the MMM more realistically estimate how marketing works in the real world by allowing:

• Brand campaigns to influence future weeks

• Prospecting campaigns to nurture awareness over time

• Direct response channels to show small, short-lived ripple effects

• Seasonal campaigns to peak and fade naturally

Importantly:
Adstock makes the model smarter, not more complicated for you.
It helps us avoid over-crediting the “wrong” week or undervaluing channels that work slowly.

How does adstock work? The intuition

Every channel has two key adstock components:

1. Lag / Peak Week (Δ) – When the impact is felt the most

2. Decay Rate – How quickly the effect fades over time

A simple mental model:

• You spend $100 in Week 1.

• Maybe you see 60% of the impact in Week 1.

• Then 25% in Week 2.

• Then 10% in Week 3.

• And so on, decaying exponentially.

The actual numbers are data-driven—which brings us to the important part…

There is no one-size-fits-all decay rate

This is where marketers often look for a simple answer—
“Just tell me: how long is the decay for Meta? What about TV? Or for my brand campaign?”

Unfortunately (and fortunately), there is no universal answer.

Decay depends on:

• Channel (e.g., TV vs. Direct Mail vs. Paid Search)

• Customer journey (fast or slow decisions?)

• Offer (a 48-hour flash sale behaves differently than a 6-month coupon)

• Creative (brand storytelling vs. direct response copy)

• Business maturity & brand equity

For example:

• A mailer with a 6-month coupon may have a long decay curve.

• A short-lived 2-week sale will spike quickly and decay almost immediately.

• Brand marketing often has meaningful ripple effects for weeks or months.

• Performance marketing (e.g., paid search) might peak the same week and decay very quickly.

This is why Paramark does not assume one decay rate for every customer or channel.
Instead, each MMM learns the shape of the adstock curve from your actual data.

General principles for thinking about decay

While every business is unique, a few principles hold almost universally:

1. Direct response channels have short decay.

Search, retargeting, and lower-funnel social typically peak quickly and fade fast.
Consumers are already close to converting.

2. Upper-funnel and brand channels have longer decay.

Awareness-building channels—TV, video, OOH, brand Prospecting—create effects that accumulate and dissipate more gradually.

3. Larger or episodic events can have long, uneven tails.

Big brand moments, sponsorships, PR events, or promotions may spike quickly but stay elevated longer (or behave in unexpected ways).

4. MMMs rarely measure beyond a year.

Beyond ~12 months, any signal becomes too diluted and too hard to isolate from other forces (competition, market shifts, brand equity, macro trends).

How Paramark models adstock

We use a flexible framework that:

• Models lag + exponential decay

• Allows channel-specific differences

• Doesn’t assume unrealistic long-term effects

• Caps practical measurement at ~13 weeks for short term channels and ~1 year for brand channels

In other words:
Your adstock curve is learned, not guessed.

This avoids oversimplification while staying grounded in what's statistically identifiable.

Common customer questions about adstock

“When should I expect to see the impact of a campaign?”

It depends—and the MMM’s learned decay gives us the best read.
Direct response campaigns often show impact quickly.
Brand campaigns take longer and peak later.

“How often do I need large brand events to maintain awareness?”

There’s no simple rule because it depends on your decay pattern.
For many clients, we recommend running experiments to compare “episodic big events” vs. “always-on smaller investments.”

“Can you show me the week-by-week waterfall of my spend?”

We can provide directional guidance, but the exact week-by-week shape is custom to the channel and the MMM’s learned parameters.

“Why can’t we just set the decay rate manually?”

You can as an experiment, but doing so introduces strong assumptions—and potentially bias.
Our philosophy is to let the model learn what the data supports.

"Your guidance is that direct response channels have short decay. But my Meta adstock shows an 8-week peak. Which is right?"

Both can be right, depending on how Meta is being used in your marketing mix.

Meta (and other platforms like LinkedIn or YouTube) is not inherently a direct response channel or an upper-funnel channel. The adstock curve reflects how the channel is actually functioning in your business. If your Meta campaigns are focused on brand awareness and prospecting, the model will learn longer decay. If they are retargeting users close to purchase, it will learn shorter decay.

The general principles in this article describe how channels tend to behave when used in a particular way, not rules that apply to a platform name. When your MMM shows a result that looks unexpected for a platform, the first question to ask is: how is this channel deployed? The answer usually explains the adstock shape.

If you are unsure how to interpret your channel's adstock curve, your Growth Advisor can walk you through what the model learned and how it maps to your actual campaign strategy.

"What does this mean for LinkedIn or other upper-funnel social campaigns when I design an experiment?"

A practical takeaway is that some upper-funnel social channels can peak later than marketers expect. For example, a B2B team may see LinkedIn's strongest SQL impact arrive roughly 3 to 3.5 weeks after impression, with meaningful carryover lasting up to about 4.5 to 5 weeks.

When that happens, a very short test window can understate the channel's impact because you stop measurement before the effect has time to show up. As a rule of thumb, if you are testing a channel with this kind of lag, plan for a long enough run window and a cooldown period so late conversions can be observed. In practice, that often means running the experiment for at least 6 weeks and allowing roughly 2 additional weeks of cooldown before final interpretation.

The exact timing will still vary by campaign objective, audience, creative, and sales cycle. The important principle is not that every LinkedIn campaign has the same decay curve, but that experiment duration should match the lag structure your channel is likely to have.

“Can I compare my MMM adstock curve to my MTA lookback window?”

Yes—but only as a directional comparison, not a one-to-one reconciliation.

Your MTA lookback window is a reporting rule: it defines how far back your attribution tool will look when assigning conversion credit to an observed touchpoint. Your MMM adstock curve is a modeled estimate of when spend is likely creating incremental impact over time. Those are related concepts, but they are not measuring the same thing.

A useful way to compare them is to ask two questions:

1. When does this channel appear to peak in the MMM? That tells you when the channel's strongest modeled impact tends to show up.

2. How long does meaningful carryover persist? That tells you whether a very short MTA lookback window may under-credit slower channels.

For example, if your MMM suggests an upper-funnel channel has a multi-week tail, but your MTA only uses a 7-day lookback window, your attribution system may miss part of that channel's delayed effect. On the other hand, if a lower-funnel channel peaks quickly and decays quickly, a long lookback window may give it more credit than is truly incremental.

The right takeaway is usually not “make the windows match exactly.” It is: use MTA for short-window touchpoint reporting and MMM for understanding the true time profile of incremental impact. If the two tell materially different stories, that is often a sign to validate with an experiment or ask your Growth Advisor to walk through the learned adstock shape for that channel.

So how should you use adstock as a marketer?

Adstock isn’t something you “optimize” directly.
Instead, it gives you a more realistic understanding of how your spend works over time—and helps you make smarter decisions, such as:

• Understanding why a channel’s performance looks delayed

• Evaluating brand marketing’s impact more fairly

• Designing experiments with enough run-time

• Avoiding overreaction to week-to-week fluctuations

• Planning budgets with an understanding of lagged effects

In short, adstock helps connect the science of MMM with the reality of human behavior.

Final Takeaway

Adstock is about acknowledging a simple truth:
Marketing echoes. The echo fades—but not instantly.

Every business has its own echo pattern, and that pattern tells us a lot about how to invest, where to test, and how to interpret results.

Our job is to help you understand that echo—not oversimplify it—and use it to make your marketing more effective over time.

Want to learn more about marketing mix modeling? Check out other Paramark marketing mix modeling articles here