EXPANDING OGILVY'S 'INTENTION X IMPACT'...

Given the glut of frameworks bobbing in the PowerPoint ocean of consulting—business, brand, consumer, communication, or otherwise—strategic approaches or thought processes are conspicuous by their presence. Yet, they can be served on tap as part of a delightful smorgasbord by strategists/planners, looking not just to juxtapose ideas and data points curiously but also to indulge themselves as practitioners of what is infamously branded as a nerdy club of theoretical blasphemy! And, in such a buffet spread, 奥美 's 'Intention X Impact' framework is pretty delectable. Subramanian Krishnan brought it up earlier this week. He spoke about value being some factor of 'unreasonable irrationality,' equating it to creativity. When we try to make sense of this interesting definition through the machinations of this framework, we notice an epiphany-'landmass,' nestled between the 'inspiration' and 'perspiration' squares.

Almost all of us would classify epiphanies in advertising or similar creative trades as fireflies. We will likely see them as plays, moments, and/or shifts of Brownian motion brilliance that impact 'IMPACT' dramatically but are too expensive to quantify!

While I appreciate the creative principle, what sets this framework apart is its ability to provide mathematical and statistical tools for predicting and quantifying value in advertising. This predictive aspect, albeit probabilistic, brings a sense of certainty to the table.

The 'Intention X Impact' framework, as presented by Ogilvy, is depicted as a linear plot. This visual representation clearly shows a linear vector , symmetrically intersecting the 'perspiration' square, providing a structured view of the framework. However, real-world scenarios very rarely trace such a trajectory. If we analyze hundreds of datasets that revolve around consumer behavior, we will notice that both 'Intention' and 'Impact' are non-linear. They are far more likely to follow other curves that could predict epiphanies better, leading to a sharper assessment of 'value'.

Before embarking on this exercise, let us outline the current assumptions within this framework.

• Linear Vector: The assumption that progress is linear (from efficiency to epiphany) oversimplifies the relationship. It assumes that impact grows uniformly with intention, which is rarely true.
• 2D Representation: By restricting this to two dimensions (Intention and Impact), the framework currently omits other crucial factors that could affect the value, such as risk, resource constraints, or randomness.
• No Consideration for Probability: The chart doesn't account for probabilistic uncertainty. At the origin (where impact is incremental and intention is low), decisions often involve significant uncertainty, where various outcomes are possible, depending on how efficiently one manages risk.

A more complex (but accurate) alternative would begin with the following alterations...

• Non-Linear Vector: The progression from "Perspiration" to "Inspiration" could be nonlinear. In reality, effectiveness and value creation often accelerate (exponentially) or decelerate (logarithmically) based on effort, breakthroughs, and feedback loops. Thus, the vector would follow a curved trajectory.
• 3D Representation: Adding a third axis could introduce a new factor, such as Risk/Probability, representing the likelihood of achieving the desired outcome. This could visualize how both intention and impact interact under conditions of uncertainty.
• Probabilistic Approach: The area near the origin (where both intention and impact are low) should include probabilistic variance. Here, even a tiny increment in intention could lead to a giant leap in effects (if you get lucky) or no change, depending on chance. In mathematical terms, we can represent this uncertainty using probability distributions (e.g., normal, binomial).

Thus, an alternative representation would have the following:

• X-axis: Impact (Incremental to Unreasonable).
• Y-axis: Intention (How to Why).
• Z-axis: Risk/Uncertainty (Low to High).

This means that we are now looking at sigmoid or logarithmic curves (or both) to account for diminishing (damped) returns or emerging epiphanies, respectively (with sustained campaign spending!).

We will, therefore, model the above as a parametric 3D curve.

Where:

• A, B, C, D are constants that scale the axes appropriately.
• t is the effort or time invested.
• f(t) represents the non-linear growth of impact over time.
• g(t) shows how intention shifts from how to why (e.g., a decaying exponential that eventually converges).
• h(t) models risk or uncertainty using a sigmoid function, which gradually rises and stabilizes.

MODEL INTUITION

• Early Stages (Low t): Efforts in the early stages might translate into incremental impact, but there is high uncertainty in the perspiration zone. Intention tends to start low.
• Mid-Stages: As intention aligns more with 'why you do things,' the impact grows, and uncertainty may decrease as things become more precise or manageable.
• Later Stages (High t): You reach the inspiration zone, where epiphanies occur, leading to an unreasonable impact. By this time, intention has fully aligned with purpose, and risk is either low or manageable.

WHAT WE CAN LEARN FROM THIS FRAMEWORK ALTERATION --> MODEL

• Early on: Progress is slow, with high risk and modest impact. This is the phase where we are still figuring out how to optimize processes and align with purpose.
• Midway: Intention becomes more aligned with "Why." Impact accelerates, and risk decreases.
• High stage: Impact reaches "unreasonable" levels, accompanied by a clearer sense of purpose and reduced uncertainty.

However, there is still headroom for model-strategic refinement because Intention is not necessarily a one-stop shop. Intent can change over time as it navigates a period of shifting purpose (from a brand perspective) or shifting focus (from a business or organizational standpoint). This means that Intention is more likely to be unpredictable in the real world. Thus, we might not want to etch a smooth curve for Intention but instead introduce discrete jumps. If we were to double-click/pinch zoom our intention curve, therefore, we'd realize that it is a combination of different phases, each of which has different trajectories.

SECONDLY, not all unreasonable impacts are 'unreasonable' or disproportionate. Impact can ebb and flow based on multiple factors that could come into play during a brand campaign or calendar. This curve modification would involve both impact acceleration and saturation.

THIRDLY, Risk reacts to shifting contexts such as market volatility, dynamic competitive landscape, etc. Therefore, risk can be a time-variant dimension. From a curve standpoint, the risk curve must either be stochastic or have a noise factor!

Also, how about introducing yet another dimension to this framework? This dimension can be resources or constraints such as fluctuating ad spending or talent deployed.

Lastly, we will need to 'randomize' the impact, considering that outcomes can be unpredictable. We can model impact using a Gaussian distribution with added variance for this.

WHAT THESE MODEL REFINEMENTS SHOW US:

• Realistic Variability: real-world scenarios where projects don't always have smooth, linear progressions. Impact and risk fluctuate, making outcomes less confident and more dynamic.
• Risk Management: The fluctuating risk curve highlights that uncertainty can arise at any point even after reducing risk with careful planning.
• Impact Surprises: The stochastic nature of impact reveals that some efforts yield unexpectedly large results, while others might fail to deliver.

DATA AGGREGATION, TAXONOMY, ENGINEERING FOR DASHBOARDED VISUAL AND AND PREDICTIVE ANALYTICS (ASSUMING A BRAND BUILDING SCENARIO OVER TIME AND MULTIPLE CAMPAIGNS):

Suggestive Key Variables:

Campaign Intention (Y-axis):

• Marketing Objective: (Awareness, Engagement, Conversion, Loyalty)
• Target Audience: (Demographic details, Psychographic segmentation)
• Message Relevance: (How well the message aligns with the brand purpose - can be a Likert scale)
• Content Quality Score: (Qualitative analysis of campaign creative and messaging alignment with brand purpose)

Campaign Impact (X-axis):

• Reach & Impressions: Number of people who were exposed to the campaign.
• Engagement Metrics: Likes, shares, comments, click-through rates (CTR), video views, etc.
• Conversion Metrics: Sales, leads, sign-ups, etc.
• Sentiment Analysis: Social media and feedback sentiment, brand mentions, and net promoter score (NPS).
• Brand Lift: Awareness, recall, favorability (before and after campaign surveys).
• Sales Lift: Impact of campaign on sales revenue or units sold.

Campaign Risk/Uncertainty (Z-axis):

• Budget Deviation: How closely the campaign is sticking to the budget (over or under-spend).
• External Factors: Market conditions, economic trends, competition, unexpected events (e.g., PR crisis, regulatory changes).
• Forecast Accuracy: Difference between predicted and actual performance metrics.
• Competitor Activity: New campaigns, product launches, or market share fluctuations that could affect campaign outcomes.

Dashboard Insights:

• Impact Curves: By plotting the Impact metrics (Reach, Engagement, Conversion, Brand Lift, Sales Lift) over time, we can visualize how the campaign's impact evolves and correlates with intention.
• Intention Curves: We can track how the content quality and message alignment evolve. For example, campaigns with higher quality scores may show a more substantial impact.
• Risk Curves: We can visualize risk factors (budget deviation, external market conditions, competitor activity) to see how uncertainty influences the outcome. High-risk campaigns may produce unpredictable results, which could either lead to breakthrough success or poor performance.
• Probabilistic Visualization: By adding variance to impact projections, you can create shaded areas or error bars on the impact curve to represent uncertainty in outcomes, especially for high-risk campaigns.

FINALLY, Integrating Data in a Dashboard:

Real-time pipelining from various sources (all-party data) can help us run this dynamic dashboard, and we can visualize it.

• Dynamic Curves: Tracking each campaign’s progress along the X (Impact), Y (Intention), and Z (Risk) axes in real time.
• Risk Indicators: Using color coding (e.g., red for high risk, green for low risk) to highlight campaigns that require more attention.
• Predictive Analysis: Using historical data to predict future trends in impact or risk and overlay these predictions onto the curves for decision-making.

Alternatively, we can use Machine Learning models to carry out the same exercise and bring this abridged 'Intention X Impact X Risk' framework to life.