February 17, 2020


Forecasting is one of those things we’re expected to do as marketers. We fundamentally understand its importance, and we’re probably all doing it—but with varying degrees of success.

Good news. Those of us charged with sharing predictions tied directly to revenue performance—can breathe a collective sigh of relief. The forecasting tools at hand have never been easier to use or more adept at predicting the future with accuracy to help us strategize, plan and react in a timely way.

If you’re not using a formalized forecasting process, or your projected revenue doesn’t consider day of year or day of week seasonality, you will undoubtedly find yourself at the mercy of unpredictable outcomes. But if you make strategic and operational decisions based on accurately predicted performance, you will set yourself up to successfully hit your goals.

Why Forecasting is Important

Done right, forecasting allows companies to make proactive and timely decisions to address shortfalls or to maximize on opportunities. There are a variety of ways to tackle forecasting. Today, machine learning and advanced statistical approaches are the most accurate.

conDati recently hosted a webinar, presented by Iris Lieuw, VP of data science at conDati, and me—on this topic that showcased marketing revenue forecasting using conDati RevenueLift.™ Here are some of the webinar’s key takeaways.

telescopeConfidently Plan and Meet Your Numbers

The beauty of forecasting is that it helps marketing leaders drive consistent results.

  • Reduces unnecessary spending. Plan according to seasonality to shift spend during expected low periods.
  • Informs campaign and resource scheduling. Lock in agencies and outside resources to meet deliverables, allowing enough lead time to capitalize on peak demand periods.
  • Eliminates missed opportunities. Understand your expected revenue rhythm to identify opportunistic periods where you can invest more.
  • Helps you budget your marketing spend early. Forecasting a year ahead makes budget allocation across channels and campaigns much easier.
  • Gives you a target to compare your progress against. Regularly monitor and compare actuals against predictions, so you know if you’re on track or not.

Approaches to Forecasting

There are different approaches to forecasting, with varying levels of accuracy.

Based on Historical
Conversion Rates



Forecasting Using
Machine Learning
  • Basic approach
  • Uses historical rates of conversions between tracked metrics and revenue
  • Conversion rates need to be reasonably consistent over time

The “Wow” Factor Is in the ML

  • You’ll get automated reporting at your fingertips. This gives stakeholders access to the information, as often as they like, eliminating the need for time-consuming ad-hoc reporting. Forecasting in Excel just does not have the same power and insight.
  • You can align sales, marketing and finance around one source of the truth. No more hassles with different functional groups reporting different numbers and forecasts. Keep the focus on driving toward business goals and away from whose numbers are right.
  • You’ll get a forecast that is form-fitting to your specific business. An advanced statistical approach is taken that uses historical data to understand the seasonality and trends of a company—taking regressors or potential reasons for variation into consideration.
  • You’ll be alerted to anomalies that fall outside your range of normal. Allowing marketers to make proactive, timely decisions to address shortfalls or maximize on opportunities.
  • You can better understand your revenue rhythm. Predicting the peaks and valleys of your revenue 12 months out gives you valuable insights into customer behavior, and helps you plan and schedule much more effectively.
  • Forecasting models built using machine learning are the most accurate. Models built on ML use a multi-pronged trend, seasonality and regressor approach that is highly accurate and explainable. Marketers making key strategic and operational decisions based on forecasts are set up for the best possible outcomes.
what-it-takes-to-accurately-forecast-revenueWatch the ''What it Takes to Accurately Forecast Revenue ... conDati Shows You How' recorded webinar, to hear conDati VP of Revenue Marketing, Kelly McKeown and VP of Data Science, Iris Lieuw, discuss how to accurately forecast marketing revenue using machine learning.
Watch the webinar replay

With accurate forecasts in hand, be ready to act on the data

Once you have your forecast in place, it’s important to establish an action plan in case results dip below where you want them to be. Knowing ahead of time that you’re going to have a shortfall is great but scrambling to understand why, what to do about it, and whose responsibility it is—puts unnecessary stress on everyone involved. Being prepared helps with organizational agility and gets things back on track faster.

The process is not complicated at all. I recommend these three steps:

  • Regularly monitor actuals against your forecast and set up stakeholder alerts should metrics fall outside of pre-determined thresholds.
  • Establish a process to diagnose the underlying drivers of anomalies. This should include identifying the team or person who will be responsible for this.
  • Have an action plan in place with strategies and tactics to address shortfalls, so you can get performance right back on track.

With the availability of tools like conDati RevenueLift, you can make strategic and operational decisions based on accurately predicted performance. A far better scenario than finding yourself at the mercy of unpredictable performance outcomes.

If you have questions or just want to see hoe predictive forecasting works first hand Schedule a demo of RevenueLift™ Forecasting today.

To watch the full 22-minute webinar “What it Takes to Accurately Forecast Revenue ... conDati Shows You How” highlighted in this blog, access it here.