The 4 Types of Data You Need for Simulation Forecasting
“I don’t have enough data.” “I don’t have good enough data.” “My data is too old.” These are the responses we often hear when customers are trying to improve their analytics processes. We hear these data concerns from small start-ups to Fortune 500 companies. In virtually every case these concerns may be true if you are using traditional models. However, simulation opens the door to a new way to use data.
When we explain how simulation works to prospective customers, they are excited at the possibility that products like ours, Concentric Market®, can help answer their business questions. By unifying analytics, building forecasts, and accelerating analytic processes, we help people answer these questions:
- How do I maximize revenue and demonstrate ROI?
- How do I create new product strategies that will lead to success?
- How do I identify the optimal marketing mix?
This bold promise flies in the face of a world dominated by data management marketers who hoodwink people into believing that perfect data is needed.
Accordingly, many people quickly become concerned about whether or not they have sufficient data to address those questions accurately. They become self-conscious about the state of their data. Much like a teenager’s body image – it’s never good enough.
So we set out to right this wrong. We did this by using different science. Simulation is the answer – almost an antidote for time series analysis. One of the benefits of using simulation software instead of time-series analysis is that you are not constrained by the hard data points you have for every input to produce an informative forecast that can be used to make data-driven decisions. Simulation allows you to use both qualitative and quantitative intelligence to build your simulation and forecast. For example, you may not have consumer research data points for awareness or consideration levels, but as an expert in your field you know how you compare to the competition. Simulation allows you to use your human intelligence to make assumptions, that are later validated for accuracy with observable outcomes like sales results and brand metrics.
Building a simulation is an iterative process. Just because you don’t have the data today doesn’t mean you should wait to start building models in order to inform strategic decisions. Even with limited data, you can still produce a forecast on day one to answer your high-level business questions. The accuracy of those forecasts will improve over time as more data becomes available and you get more confirmatory outcomes in the market. As you integrate more data sources and validate observations, you can expand the model to inform more nuanced and detailed what-if questions.
There are four main buckets of information that users typically provide to make a simulation work, in virtually every organization we serve, our customers have enough collective intelligence within the organization to create a simulation. By combining concrete datasets with employee or team intelligence that may be more qualitative in nature, you are able to build a framework of your market.
The first set of information that is used to define the market is the outcome you are trying to influence. Most often, company sales data is used here. Concentric Market® and other simulations can model more than sales. They simulate a variety of human behaviors such as, website visits, online application submissions, or program enrollments. It is helpful to have outcomes by segment, population size by segment, and purchasing power by segment. But you don’t need this level of detail to get started. You can continue to integrate new data into the model to make it more comprehensive over time.
Since you are simulating an entire market – meaning how people within a certain category responds to the choices available to them – you will also need sales and marketing strategies for your competitors. Competitor data may be available from syndicated sources, but you can use approximate information that is available from public sources, annual reports, or analyses from business experts.
The third area of intelligence needed is around your customer. Concentric Market® simulates an audience’s likelihood to choose one option instead of another. You will need information around how they think. This includes consideration (or awareness), perceptions, and the relative importance of the attributes in driving a decision. We often find these data sets are already being collected through brand health studies. Even if there isn’t data from a quantitative study, your brand experts may use their judgment to make initial estimates of these values. These initial estimates will be verified later through calibration and forecasting.
The fourth bucket of information integrates the marketing activity that was in the market to influence a person. This can be as simple as marketing investments broken out by paid, owned, and earned or it can be as granular as the tactics or even specific media outlets within each discipline. Unlike marketing mix analysis, Concentric Market® accounts for how activities like word-of-mouth, social media and PR influence business outcomes in addition to advertising and in-store activities.
Use the Data You Have To Get Answers
We have customers who have successfully built models relying on qualitative expertise and judgment while others have leverage quantitative data spanning many years. In either case, we have had customers achieve very accurate forecasting power with an error rate under 10%. The breadth and depth of the data may play a role in determining the types of questions that are addressed. Concentric Market® is flexible and helps organizations make the most of the intelligence they have today.
If you still have doubts about your data after reading this post and would like to learn more, our team is here to advise you on how to increase the accuracy of your modeling and forecasting efforts with Concentric Market®.
You can learn more about the Concentric model here.
John directs training on Concentric Market and provides guidance to our users on setting up market simulations, using various data sources, validating simulation models, and applying simulation to answer key questions. With a background in engineering and quantitative risk modeling, he provides analytics insight to our users through direct interactions with our customers and efforts to continuously improve Concentric Market algorithms.