How to forecast marketing results when you don’t have historical data

Marketing forecasting shows which channels, tactics, or campaigns are most likely to deliver results. But when you don’t have much historical data, even the best forecasting models fall short.

Maybe you’re entering a new market, testing a new channel, or pitching a strategy no one has tried before. Without past results, it’s difficult to estimate performance, set expectations, or justify budget decisions.

That’s why you need a structured way to forecast marketing with limited data. Advanced planning platforms now allow you to market forecast without historical results by combining proxy data, category benchmarks, and elasticity patterns from comparable markets. These inputs model how changes in spend, mix, or timing could affect conversions, revenue, or margin.

In this guide, we’ll break down how to forecast marketing results with minimal data, the methods of marketing forecasting, and the tools that make it possible.

What is a marketing forecast?

A marketing forecast is a way to use data to estimate future marketing results. This method looks at past performance, market trends, and customer behavior to predict what may happen next.

Forecast analysis lets marketers estimate likely outcomes—such as leads, conversion rates, and revenue—before allocating budget. It also shows how different strategies can improve acquisition, retention, and growth.

Why do you need marketing planning and forecasting?

You need marketing planning and forecasting because they show where to invest and what results to expect from each channel and budget scenario. Instead of reacting to outcomes after campaigns run, you can plan with data and anticipate performance before you spend. Forecast marketing is especially critical for long-term strategies, such as annual planning, which provides the data-backed direction and measurable goals necessary to align resources and budgets over a full year.

With this approach, marketing teams can:

• Connect strategy to measurable marketing impact. When you see how each campaign drives growth and revenue, your next moves become clearer.
• Identify drivers of results. By knowing which channels perform best, you can invest where impact is strongest.
• Adjust quickly to changing conditions. Use real-time data to shift budgets and tactics before performance drops.
• Set clear goals and timelines. This step helps your team stay aligned and set realistic expectations for ROI.
• Reduce uncertainty. When you replace assumptions with a data-driven marketing approach, every decision becomes more confident and grounded.

An effective marketing planning process sets priorities and allocates resources, while forecasting validates these choices and adjusts strategies over time. This combination helps teams act with precision and adapt to change.

Learn more: How forecasting helps brands drive optimal profitability.

Types of data needed for market forecasting

Marketing forecasting starts with correct data. But many marketers still struggle to collect, clean, and transform that data into meaningful insights. Here’s what you’ll need:

1. Accurate data. If your data is incomplete or outdated, your projections will mislead you. Collect verified data from CRM systems, web analytics, ad platforms, and revenue reports. 
2. Historical data. You need to understand how marketing efforts are performed over time, so you can spot seasonality and performance trends that drive realistic forecasts.
3. Market size. That’s what defines your growth potential. Estimate your total addressable market (TAM) and the portion you can realistically reach. Use industry reports, pricing benchmarks, and audience data to calculate revenue potential.
4. Target audience. Know your customer before you forecast marketing. Precise projections depend on understanding who buys, why they buy, and when. 
5. Analysis of external factors. Even the most accurate marketing forecast can fail when external conditions shift. Economic trends, competitor activity, and cultural events all shape marketing performance.

Keep learning: Marketing industry trends insights

5 common marketing forecasting methods 

There are several methods you can use to predict marketing outcomes. Your goals, data quality, and market context determine which approach is most effective. Common types of forecasting in marketing include:

1. Predictive modeling 

Predictive modeling uses statistical and machine learning methods to forecast marketing results from historical data. It enables scenario-based planning, so teams can test “what-if” scenarios, compare performance over channels, and predict ROI before spending.

You can use predictive modeling in marketing to:

• Simulate budget shifts and campaign outcomes before launch
• Identify which channels or tactics drive incremental growth
• Estimate future ROI with greater accuracy
• Replace guesswork and marketing instincts with data-backed planning

Keep learning: Why scenario planning is the evolution in marketing forecasts

2. Delphi technique

The Delphi technique gathers insights from a panel of experts through multiple rounds of questioning and feedback. Each round refines their opinions until the group reaches a well-informed consensus. 

With the Delphi technique, you can:

• Forecast in new or emerging markets with little historical data
• Validate assumptions through collective expert insights
• Build confidence around strategic decisions when numbers alone can’t

3. Correlation technique

The correlation technique examines how different variables move together. For example, how does ad spend impact conversions, or how does engagement influence brand awareness? Correlation, however, does not establish cause and effect. 

Apply this marketing forecasting technique to:

• Spot early performance indicators or leading trends
• Explore how marketing activities relate to business results
• Combine with other models to confirm proper drivers of impact

4. Time series technique

Time series forecasting looks at how marketing performance changes over time to project what’s likely to happen next. This technique captures seasonality, recurring cycles, and long-term trends, so you can:

• Predict seasonal demand and recurring campaign performance
• Monitor how trends accelerate or slow over time
• Inform revenue projections and long-term growth planning

5. Response model technique

A response model predicts how your customers react to specific marketing actions, such as price changes, promotions, or creative variations. This forecasting method helps to:

• Test and compare campaign offers or messages
• Predict customer reactions before launch
• Optimize targeting, pricing, and creative decisions

How to forecast marketing results

Building a marketing forecast starts with structure. Before diving into tools or models, you need a clear process to organize your data, set your goals, and test assumptions. Follow these eight steps to turn raw information into forecasts you can trust.

1. Define your goals: Start by clarifying what you want to predict. For example, website traffic, conversions, or marketing-driven revenue.
2. Gather and clean your data: Accurate forecasts depend on solid data. Collect metrics from your CRM, analytics tools, and campaign platforms. Check that your data is complete, consistent, and up to date before you start modeling.
3. Analyze customer and market trends: Internal data alone doesn’t give you the whole picture. Add context by looking at market trends, seasonal patterns, and competitor performance.
4. Identify key marketing performance indicators: Focus on the metrics that matter most to your goals. For example, cost per lead, conversion rate, or customer lifetime value. Tracking too many numbers can make your forecast messy and less useful.
5. Model your sales and marketing funnel: Map how leads move through your funnel, from awareness to conversion. Use historical data to estimate how many are likely to convert at each stage.
6. Incorporate predictive and scenario modeling: Test different “what-if” scenarios through AI-driven or statistical models. For example, simulate how changes in ad spend, creative strategy, or pricing might affect your conversion rate or ROI.
7. Allocate budget and resources: Once you’ve made your projections, plan your resources accordingly. Invest more in the channels that deliver the best results. Keep some flexibility to adapt to new trends or performance shifts.
8. Review and adjust regularly: Your forecast should evolve over time. Compare predicted results with actual outcomes, analyze the gaps, and update your assumptions. 

How to make marketing predictions if you don’t have data

It’s now possible to make marketing predictions with no data following structured inputs that estimate performance, such as:

• Proxy data: Data from similar campaigns, products, or audiences that stand in for missing historical records. Proxy data estimates outcomes when performance records are unavailable. For example, results from a similar product launch in another region forecast your own.
• Category benchmarks: Industry-standard metrics or averages that indicate expected performance. Category benchmarks serve as a reference point for metrics such as click-through rates, conversion rates, and ROI. For example, the average engagement rate in your product category sets realistic goals.
• Elasticity patterns from comparable markets: Patterns showing how changes in spend or other variables historically affected results in similar markets. Keen’s marketing elasticity engine, for example, you can predict how your campaigns might respond to budget shifts or new strategies.

Modern marketing mix modeling (MMM) platforms, such as Keen, apply these inputs to predict the marketing impact with minimal data. Our marketing elasticity engine leverages billions of dollars of aggregated media spend data and thousands of performance patterns to simulate thousands of possible scenarios rapidly.

Basic descriptors such as business scope, target audience, and primary marketing tactics run the simulation. Detailed client spend history is not required. With minimal input, Keen’s Planning Module generates probabilistic forecasts and optimized budget recommendations.

Get the best marketing forecasting tools with Keen

Traditional marketing forecasting tools rely on past ROIs and industry best practices, such as standard playbooks based on category benchmarks and historical trends; however, these references miss how channels actually interact. That’s a problem we can solve with our Planning Module.

To do that, we don’t require a large dataset. Instead, we allow you to tap into our marketing network, which we call our core asset. The Planning Module is designed to work with minimal inputs that are easily accessible or estimated, even during a pitch. The average time to complete a plan is just five minutes.

For example, when setting up a plan for a major fitness company, we only need a few high-level descriptors: 

• Business scope: Product, geography, and audience (for example, stationary bikes/treadmills, the United States, at-home gym rats)
• Core metrics: Desired output (for example, sales units) and basic financials like annual volume, average price, and margin (for example, 1 million units, $2,000/unit, 40% margin)
• Tactics: Relevant channels from our library of 200 different tactics (for example, TV, Meta, Google Search)

Once the inputs are finalized, the module runs 1,000 simulations, and our engine factors in the interaction effects, seasonality, and differing market conditions across all possible investments.

The output delivers a risk-adjusted forecast and a clear plan on how to achieve your goals:

• Max profit investment: The precise point of diminishing returns, showing the optimal marketing spend. For the fitness company, it was $135.2 million annually

• Optimal allocation: A clear breakdown of where to allocate budget for scale, often identifying channels like TV and Meta as the “lion’s share” for a responsive brand like this fitness company
• Risk curves: A transparent probabilistic forecast, presenting a probabilistic forecast that shows the potential downside and upside of your strategy, avoiding “false confidence”

With Keen’s Planning Module, you move from guesswork to a data-driven strategy, uncovering optimal investments that are simply impossible to identify manually due to the staggering number of variables. In the words of Josh Lucas, Keen’s Head of Product Development:

“If you have 5.2 billion different combinatorial possibilities, odds are you’re probably doing the same thing and you’re not getting the full picture, right? That’s one of the biggest things Keen brings here. A perspective that only a machine can give you that a human could never have.” 

Request a Keen demo to see our Planning Module in action.