The Tidal Wave Model

Marketing Today is Inherently Reactive

The current marketing landscape relies on hindsight to make decisions for future campaigns. Outside of one-off large experiential executions, an overarching sentiment of ‘if it’s not broke, don’t fix it’ exists. What this means is that marketers typically will wait for a campaign to wrap, see what worked using reporting insights, such as Media Mix Modeling (MMM) or Multi Touch Attribution (MTA) and then incorporate these findings into their next campaign. More recently, marketers also use real-time analytics and optimization, which similarly require a campaign to already be live to make changes. The foundation of the industry is therefore inherently reactive, which affects campaign efficiency and performance, leading to overspending.

Here's the problem: post-campaign findings on what drove attribution don’t give you a full picture. If podcast ads drove the majority of conversions last quarter, should marketers increase their investment there? What if the audience is already saturated, or increased spend triggers audience fatigue in the next cycle? These retrospective analyses highlight past successes but don't effectively predict or model future performance. It's like trying to steer a ship by looking at its wake.

The Marketing Funnel is Outdated

This reliance on retrospective and linear assumptions underscores a deeper issue in marketing: an oversimplified view of consumer behavior. Traditionally, marketers have visualized the consumer journey using the Marketing Funnel, where consumers trickle from one stage to the next (Awareness → Consideration → Conversion → Loyalty) in a linear fashion, with the number of users dwindling down each stage. However, recent research from Google and BCG has demonstrated that real consumer journeys aren’t linear at all. Shoppers loop, revisit, compare, stall, and accelerate before making a purchase. Instead of a linear journey, we’ve created a new framework to reflect this reality: The Tidal Wave.

Attention as a Metric

Advertisers are continuously fighting for consumers’ mental availability. In the past, they’ve leveraged tactics such as jingles and mascots to stay memorable to people. Nowadays, this has become even more difficult; in today’s omnichannel world, ad exposures are constant: streaming video, social feeds, digital out‑of‑home, podcasts and more serve ads around the clock. To better compete among this crowd, think of ad exposures less as an isolated ‘blast’ or impression, and instead think of it as a wave. The wave rises when you are exposed to an ad, and fades as you forget it. In order to increase mental availability, the goal is to expose someone to another ad before the effects of the first one completely wane. Thereby, you are able to build a larger wave and take up more mental space.

Why Use a Wave?

With more mental space, comes more of a mulling over by the consumer, what Google describes as the ‘Messy Middle’, a timeframe wherein people will waver between exploration and evaluation of the product. Encouraged by repeated ad exposures reminding them of the product, a consumer will make the decision to ‘convert’, taking the action suggested in the ads.

To best model these waves, we leverage a Gamma Kernel (seen below). This rises to a peak then falls over days, fully capturing the cognitive buildup and decay of each impression.

Wave Strengths: Active vs. Passive Impressions

As previously said, ad exposure has become something of a constant stream people are exposed to. This leads to some ads being actively engaged with, while others are only half noticed. When an ad is actively consumed, we measure the impression ‘wave’ as larger than if the ad is passively noted. Thereby, our model weights impressions by attention quality, not just counting eyeballs but measuring impact.

Active Impression Examples (full attention / intent):

• Tapping “Play” on a podcast episode and hearing the host’s ad read.

• Clicking on a paid search result and spending time on the landing page.

Passive Impression Examples (background exposure / low attention):

• Skipping past an Instagram Story ad without sound.

• Glancing at a roadside billboard while driving.

Setting Thresholds

Research shows us that, while the Marketing Funnel is outdated, the conception of phases is reflected in consumer behavior. That means, a certain number of exposures to an ad does show to lead to certain behaviors. We’ve set these thresholds as below. As an example, based on research from Google and BCG, we know that it usually takes 6–9 active exposures to drive conversion. As each wave builds on each other, they add up to reach each threshold with the active ones building higher waves (weighted at 1x bump) than the passive ones (weighted at 0.3x bump).

Awareness (1× peak)

Google research on ad recall shows that a single “full bump” (i.e. one highly attended exposure) is often enough to register brand awareness.

Messy Middle / Exploration & Evaluation (2×-7x peaks)

Studies of early evaluation behavior (e.g. Google’s “Messy Middle”) find that roughly two strong exposures get a shopper from “I’ve heard of it” to “I’m thinking about it.” Thereafter, because every old exposure decays in one’s memory and repeated exposures fatigue people, the awareness around a product can drip or stall, even as a campaign continues. That means, someone might cross the ‘Evaluation’ threshold and then drop back into ‘Exploration’, then climb back up etc. Our model therefore encompasses the ping-pong behavior Google describes.

Conversion (7× peak)

Both Google and BCG data converge around 6–8 fully attended exposures needed to actually drive a purchase decision, so we land on 7 here as our sweet-spot threshold.

Care (12× peak)

BCG’s work on post-purchase engagement and “care” suggests you need another 5 – 7 exposures beyond purchase to cement loyalty, so we benchmark that at ~12 full bumps.

Our Wave-Based Model embraces that insight by replacing rigid stages with tracking a continuous “stock” of consumer attention, capturing real back‑and‑forth behavior, and predicting not just if but when people will buy and become loyal.

Optimize Frequency

Our model accounts for saturation and interactions between media channels to determine the ideal frequency caps for a campaign. By accounting for diminishing returns from repeated exposures and accurately measuring how different channels amplify or cannibalize one another, marketers can precisely control ad frequency to maintain peak consumer engagement.

Optimize Number of Creative Assets

Inputs such as creative fatigue, recency, and creative rotation inform the optimal number and timing of creative assets. By dynamically adjusting asset weights to reflect audience fatigue and freshness, the model clearly identifies when audiences begin to lose interest, precisely signaling when to swap in new creative assets. This proactive approach prevents message burnout, preserves audience engagement, and maintains campaign effectiveness.

Optimize Flight Length

The model integrates conversion lag and network effects to identify the most effective campaign flight lengths. By accurately predicting delays between peak consumer engagement and actual conversions, and factoring in the organic amplification from satisfied customers, marketers can predict when the majority of conversions and consumer interest take place, and therefore when they should end a campaign instead of stretching it out despite diminished returns.

Optimize Investment

Through engagement convolution, our model synthesizes all inputs (frequency, creative effectiveness, timing, and cross-channel interactions) into a single, comprehensive view of campaign effectiveness. This holistic perspective ensures precise, efficient allocation of marketing budgets, directly maximizing return on ad spend (ROAS) by minimizing waste and maximizing conversions.

A Living Model that Gets Smarter Over Time

Because every campaign generates data, we layer on Bayesian updating once the flight wraps. New performance insights refine your Gamma parameters, fatigue curves, and channel weights, so the next plan begins with even more accurate projections.