Predictive research: how to use B2B market research for forecasts and strategy planning

Key takeaways: 

Predictive research helps B2B companies anticipate trends, make market forecasts, and improve decision-making. You can use historical data, statistical modelling, and machine learning to forecast market trends and assess risks. Trade-off analysis and regression modelling help test pricing strategies and sales approaches. Qualitative research explores future industry challenges through open-ended interviews, while quantitative research uses structured surveys to forecast demand. Ensuring high-quality respondents is crucial for accurate B2B market predictions.

B2B companies use predictive research to anticipate trends, make market forecasts, and react accordingly.

Market research projects can explore past, current, and future trends to help businesses grow. Let’s take a B2B energy company as an example.

An energy company studying its customers’ purchase journeys up until now can learn where there are points of friction. Improving these could help it make more sales. 

In addition, more predictive research could provide market forecasts, shaping the energy company’s future sales and growth strategies. For instance, McKinsey estimates that global spending on physical assets in the net-zero transition will reach $275 trillion by 2050, or about 7.5% of GDP.

Researchers use patterns in existing information to make informed predictions about behaviour, trends, or events: 

• On top of historical data analysis, researchers can apply statistical modelling and predictive analytics to forecast future outcomes.
• Predictive research focuses on future possibilities and causal studies examine cause-and-effect relationships with specific hypotheses.
• Machine learning algorithms improve predictive research by processing vast datasets quickly.

Businesses use predictive analysis to understand customer behaviour, forecast market trends, and identify risks or opportunities.

However, this is more challenging for B2B companies than for B2C ones. Whereas a B2C retailer could use volumetrics based on vast consumer data from thousands of purchases to optimize stock levels, many B2B companies don’t have the same level of data to use.

While there are a few challenges to overcome, the benefits of successful B2B predictive research include getting insights for:

• Long-term strategic planning
• Future-focused marketing
• Better allocation of resources
• Capitalizing on nascent opportunities
• Anticipating market risks and disruptions

Predictive research in qualitative and quantitative studies

Future-focused questions help companies identify emerging opportunities, risks, and changes in client demand. Both quantitative and qualitative research provide opportunities for this:

Qualitative research

Researchers conduct open-ended interviews to explore decision-making processes, industry challenges, and future needs. Future-focused questions reveal the reasoning behind purchasing decisions and highlight nascent trends.

Moderators ask open, future-focused questions such as: 

• “What emerging challenges do you expect to impact your business strategy?”
• “How do you think your company’s procurement priorities will change over the next five years?”

Scenario-based discussions help businesses test reactions to potential market shifts in the future. Assessing economic, regulatory, or technological changes helps companies anticipate how customers may change their spending.

While ethnographic research is rare in B2B, qualitative media diaries can provide a reliable long-term view of an audience’s engagement with marketing communications, for example.

Quantitative research

Researchers collect numerical data through structured surveys and statistical models to forecast B2B market trends.

For example, a quant questionnaire can include future-focused questions to forecast spending, assess demand for new services, and predict shifts in procurement strategies.

B2B surveys include questions such as: 

• “How likely are you to increase your budget for this type of service in the next 12 months?” 
• “Which of the following factors will most influence your purchasing decisions in the next five years?”

With a robust enough sample size, businesses can analyze the statistics from questions such as these to inform their planning decisions.

Statistical techniques and technology for predictive research

In quantitative market research, you can use advanced statistical modelling, data mining techniques, and other predictive technologies to forecast future events and scenarios:

• Pricing strategy models: Pricing research tools can forecast the impact of different price points and strategies on product uptake and revenue. Pricing techniques include: van Westendorp; Gabor-Granger; brand-price trade-off; monadic price testing; regression analysis; choice-based conjoint (CBC); and SIMALTO.
• Trade-off analysis techniques: This is a market simulation for your new products or services under a wide range of scenarios. It’s a chance to test how popular, or unpopular, they will be – depending on which features you leave out, factors you change, and what price you charge. Some of the most common trade-off approach types in research include: MaxDiff analysis; CBC; full-profile conjoint; Partial-profile and adaptive conjoint types; and self-explicated scales.
• Decision-tree algorithms: CHAID is a predictive model used to forecast scenarios and draw conclusions. It involves regression, machine learning, and branching decision trees. It’s useful for market segmentation, brand tracking studies, new product development projects, and marcomms testing.

Regression analysis features in all these examples. It can reveal the relationships between dependent and independent variables, showing what drives outcomes when research respondents do not realize it themselves.

Predicting sales, customer churn, and marketing effectiveness

Combined with a robust market segmentation, predictive analysis can help forecast sales, customer churn, and marketing effectiveness in B2B:

• Sales: Using predictive analytics on your CRM can provide valuable insights into expected customer behavior, so you can tailor sales strategies accordingly. Predictive analytics benefits include insights into sales trends, revenue forecasts, upselling opportunities, cross-selling opportunities, and more.

• Customer churn: Predictive models can analyze historical data to identify the behavioural patterns that typically precede customer churn, then warn you when a customer’s behavior is showing these patterns. You may then be able to retain the customer by taking action in good time.

• Marketing: Similarly, businesses can tailor their marketing messaging, timing, and channel usage to reach the right audience with the right message based on predicted customer behavior. This can improve conversion rates and ROI on marketing spend.

Used in combination with a segmentation, predictive analysis lets you tailor your forecasts by the different groups in your customer base. This provides more accurate results and helps you see which groups present the best opportunities for different strategies.

Other forms of B2B customer research can include elements of market forecasting too.

Thought leadership, product demand, and market forecasts

Market research projects focused on predictive modelling and forecasts include:

• Thought leadership: Businesses use thought leadership research to establish authority, predict future events, and influence decision-makers. Companies that provide accurate forecasts gain credibility, positioning themselves as trusted experts.

• Market assessment: With this analysis, you can forecast how big a market is and see if it’s growing. Sizing the opportunity often involves market intelligence research – using data to estimate the total addressable market (TAM), serviceable addressable market (SAM), and serviceable obtainable market (SOM).

• New product development research: As part of this research, you can forecast sales trends and demand for a new product by testing its concept with potential buyers. You can research the size of the opportunity by factoring in the intended price point and model information, while deducting your costs.

There are also several different business growth models you can research to forecast demand and test sales strategies.

Useful frameworks that explore potential future scenarios include: Porter’s 5 Forces, studying the threat of new entrants or substitutes, and PESTEL – evaluating opportunities by political, economic, social, technological, environmental, and legal factors.

Best practices for predictive research in B2B

#1 Experiment with generative AI, but don’t over-rely on it

Machine learning, part of the decision-tree algorithms mentioned in the previous section, is an AI field. It goes without saying that there is a role for AI in B2B research.

Currently, generative AI is having an impact on all industries. Some researchers will likely find ways to improve predictive research and deliver great results through generative AI.

However, there are risks. Large language models (LLMs) can create false information when trying to answer complex questions from a user that they do not have enough training on. 

These generative AI hallucinations are capable of fabricating statistics that sound plausible but when challenged to provide the source, sometimes the models cannot provide one.

AI in general is crucial for statistical modelling in predictive research, with many success stories demonstrated and established processes put in place. But arguably, identifying the use cases for generative AI in predictive research design is still a work-in-progress experiment due to mixed results so far.

#2 Prioritize respondent quality for accurate predictions

In market research for a B2C audience, research panels are a low-cost and efficient way to reach respondents. But most of these don’t have genuine B2B respondents, even if some claim otherwise.

For some B2B research companies, the temptation is there to prioritize quantity over quality to get more statistically robust market predictions.

But the predictions will be unreliable if they are based on respondents who don’t really reflect your typical, probably quite senior, B2B buyer. You don’t want to make big-budget decisions based on shoddy data.

Instead of taking the risk with panels, you could build a proprietary database of interested customers, using a CRM if you have clear consent combined with business analytics. Other options include finding suitable respondents using:

• LinkedIn
• Trade publications and websites
• Independent online forums
• Industry events and associations
• Emailing

Wherever you source a B2B sample, it’s vital to use several techniques designed to screen out any fake respondents. These include balanced scales, logic traps, ‘red herring’ answers, and open-ended questions requiring industry-specific knowledge.

#3 Identify the go-to-market strategy that will maximize adoption

It’s not just enough for a new product to fill a gap in the market. The business case needs to be strong and research will provide some of the key information needed to put one together.

The research needs to identify the strategic benefits of launching the product and forecast the likely demand. This information, combined with the expected costs, can help model how much profit to expect.

Even a new product that meets unmet needs – and is predicted to be an attractive commercial proposition – isn’t too big to fail after launch. To reach its full potential, it needs to go to market with a well-tested sales and marketing strategy.

Research can help identify the right sales messages and marketing channels, informing the go-to-market strategy for launching a successful product and maximizing adoption.