When we first wrote about the Six Foundations, we approached them largely as requirements for building a model properly. But if you lead a brand, the more important question is: why does this matter to me?

Each foundation exists for a business reason. In developing the Six Foundations, if we could not explain what better decision a foundation enabled and how that decision could improve a business outcome, we dropped it from consideration.

Seen this way, the Six Foundations aren’t simply characteristics of a good model. They are requirements for turning better understanding into better business performance.

1. Executive Relevance: Solve Something Worth Solving

We start with what the business is trying to accomplish, rather than with the data we happen to have or the marketing metrics we traditionally measure.

What does management want to change? Revenue? Profit? Customer growth? And what decisions could change that outcome?

This puts marketing into a more useful capital-allocation conversation. Instead of asking how much budget marketing should be allowed, we can ask:

How much should we invest given the incremental returns available?

That’s a question the CMO, CFO and CEO can answer together.

2. Holistic Design: Find All the Levers

Advertising doesn’t operate in isolation. Sales also move because of price, distribution, competitors, economic conditions, customer behaviour, geography and many other factors.

Accounting for these improves attribution. But the larger benefit is that we learn more about what actually moves the business.

One example is baseline dynamics. We can examine the underlying trajectory of the business and how it varies across markets and over time. We have seen advertising perform better in times and places where the baseline is strengthening than where it is weakening.

That gives us another lever for growth.

The objective isn’t simply better attribution. It is better understanding of the business and more ways to improve it.

3. Predictive Accuracy: Know Before You Go

Explaining what happened is useful. Predicting what will happen is much more valuable.

If management is considering another large investment, we want to predict the incremental outcome, where that investment should go, and whether another course of action would produce more.

We want to know before we go.

The more accurately we can predict the consequences of alternative actions, the less uncertainty surrounds the investment decision.

Accuracy creates confidence. Confidence moves money. Outcomes provide proof.

This is especially important when the problem is underinvestment. If management doesn’t know what another dollar will produce, being conservative is rational. Better prediction can change that calculation.

4. Actionability: Capture the Opportunity

Finding an opportunity isn’t the same as capturing it.

A model needs to operate where decisions can actually be changed — by geography, audience, channel, timing, investment level or other controllable variables.

It also needs to recognize that decisions interact. A media opportunity doesn’t have a fixed value. Its value depends partly on what else the advertiser is doing.

This means we can have individually optimized search, television and CRM programs and still have a suboptimal overall plan.

The objective isn’t to optimize every component independently. It is to find the combination of decisions that produces the best business outcome.

This also has implications for model design. We need to understand how each channel is actually planned and bought, and build the model at a level that allows the best combination of actions we identify to actually be executed.

5. Independent Testing: Earn Confidence

We shouldn’t trust a model simply because it explains historical data well.

Confidence needs to be earned.

That means testing against data the model wasn’t built on, testing predictions forward and, ultimately, comparing predicted incremental outcomes with what happens when the recommended action is actually taken.

Predict. Act. Observe. Test. Learn.

That also creates continuous improvement. Every action produces new evidence. Where prediction and outcome differ, we learn and improve.

For management, this creates a sensible path to scale: start, test, learn and put more capital behind the system as the evidence earns greater confidence.

6. Leverage: How Much Better Can We Make the Business?

This may be the most important question of the six.

Imagine an extremely accurate model that passes every test but discovers the business is already operating close to its potential.

It is a good model. But it hasn’t uncovered a particularly valuable opportunity.

Now imagine another model that identifies a substantial difference between what the organization is doing and what it could be doing.

Perhaps money is allocated to the wrong places. Geographic variation isn’t being exploited. Activities that work particularly well together aren’t coordinated. Timing could be improved.

Or perhaps the biggest opportunity is that the business should simply invest more.

That difference between the expected outcome from continuing as we are and what could be achieved through better decisions is leverage.

Marketing analytics has traditionally emphasized efficiency: How can we make the existing budget work harder?

That’s important, but it isn’t necessarily the most valuable question.

Suppose a business spends $50 million and we find a way to make that $50 million produce more. We have created value.

But suppose the evidence shows it should spend $70 million because the next $20 million can produce an attractive incremental return.

Not investing that money also leaves value on the table.

The constraint shouldn’t automatically be today’s budget.

The objective should be the best business outcome.

The Question That Ultimately Matters

The Six Foundations can therefore be reduced to six questions:

Are we solving something important?

Do we understand what is really driving the business?

Can we accurately predict what will happen?

Can we act on what we learn?

Have we independently tested our predictions?

And finally:

How much better can we make the business?

The first five establish our ability to answer the sixth.

Because the objective isn’t to build the most sophisticated model, or even simply to measure marketing more accurately.

It is to use better understanding, prediction and decision-making to find, value, prove and capture opportunities to improve the business.

That is what a Six Foundations model is designed to do.