A guide to trade-off analysis techniques in B2B research

Why use trade-off analysis techniques?

“There are no solutions; there are only trade-offs; and you try to get the best trade-off you can get, that’s all you can hope for.” 

That’s according to economist Thomas Sowell in his book A Conflict of Visions and this logic is at the heart of many business strategy decisions. Either the perfect solution isn’t possible, or it’s subjective – so, you need to make trade-offs.

A new product or service can’t do everything for everyone. It would be too expensive, or too complicated, or take too long to deliver, and so on.

There are many trade-off decisions to make. Which features should you prioritize? And how much can you charge? These are difficult questions to answer because asking customers for direct feedback using standard quantitative research won’t always work well.

If you ask them how important different features or factors are, they’ll likely say they want almost everything you can offer, for the lowest price possible, since they don’t have to make any trade-offs. If you ask them to rank criteria, that may work better, but it’s still difficult to determine the difference in impact between the winner – often, ‘price’ – and the runner-up. 

Additionally, often what customers say matters most, isn’t what they think it is, because they don’t always know precisely how and why they make their purchase decisions. They may over-rationalize the importance of certain criteria – and underestimate the importance of trade-offs.

So, what is trade-off analysis? Think of it as 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.

Trade-off analysis doesn’t only tell you the hierarchical order of features in terms of derived importance, or the most successful product package. It also tells you how much of a difference there is between the features and criteria that would drive a purchase decision.

By segmenting respondents, the analysis can also explain the trade-off decisions you should make to appeal to different customer types. Alternatively, trade-off analysis can help you identify new segments – potentially, a new subset of customers that would pay good money for your new product once it’s ready.

Trade-off analysis tools help you understand:

• Which features you must include in the final product or service
• Which are the strongest features to promote in marketing
• What price you can charge, or pricing tiers you can use
• Which features you should deprioritize
• How to differentiate from competitors
• Who to sell to

Types of trade-off analysis techniques in B2B research

There are several reasons why trade-off analysis can help you get more accurate results in B2B research.

Firstly, the purchase decision-making process is more complex in B2B than in B2C. It’s longer, it involves more stakeholders, and there’s usually more investment required. Therefore, it’s more complicated to understand the role and impact of price on a purchase decision. The statistical analysis involved in trade-off research can give a truer read on the importance of price, reducing the chance of bias or understatements.

Secondly, buyers in B2B are more senior and harder to reach than in B2C, so the sample size in research is usually smaller. By including statistical analysis, it can help to provide a more reliable read and give you greater certainty that the results are as accurate as possible.

There are different types of trade-off analysis, but they have lots in common. Overall, you start by setting out a long list of attributes or features, feature levels, or packages of features in a very specific questionnaire format. 

Next, you ask respondents a series of questions to determine their preferences. At the end of the survey, advanced statistical analysis creates a mathematical model.

These are some of the most common trade-off approach types in research:

• MaxDiff analysis
• Choice-based conjoint (CBC)
• Full-profile conjoint
• Partial-profile / adaptive conjoint types
• Self-explicated scales – quantitative or qualitative

It’s also possible to use a hybrid approach, combining more than one trade-off method type – for example, using self-explicated scales on top of a conjoint methodology.

Here are the pros and cons of the different trade-off analysis approaches: 

#1 MaxDiff analysis 

Running a MaxDiff analysis will reveal your customers’ hierarchy of relative preferences, based on an exercise where respondents make trade-offs. It is a type of best-worst scaling, although there are other forms of best-worst analysis too.

Specifically, they pick the best and the worst – or most preferred and least preferred – factors from a series of short sets. Each one only shows a few factors at a time, taken via a statistical algorithm from a long list.

By making a series of trade-offs out of different sets of factors, over time a picture emerges of a respondent’s most important priorities, as well as their relatively less important ones. The results will show numbered scores for each factor – the higher the score, the more important the factor.

In this way, the exercise reflects comparisons your customers will make when researching products or going through a decision-making process. MaxDiff analysis is a popular technique to understand product or supplier choices.

Another benefit of using MaxDiff is it lends itself well to running an additional TURF (Total Unduplicated Reach and Frequency) analysis once you have the data. It uses calculations based on either respondents’ top choice of different factors, or weighted probabilities.

Taking this extra step reveals how to maximize the reach of your product or service, by offering the best combination of features from your long list. It can also reveal how many and which features you need for the best ROI, assuming that each additional feature requires more investment. 

For example, if five crucial features together will reach most people, and adding a sixth could reach slightly more but require significant investment, TURF analysis gives you the statistical modeling to inform your trade-off decision. 

#2 Choice-based conjoint (CBC)

A conjoint analysis measures factors’ value according to your customers. It’s a regression model, so it examines the influence of independent variables on a dependent variable.

In a CBC study, respondents choose their preferred full-profile concept out of a selection, rather than rating or ranking them manually. CBC also includes an option for none of these – in contrast to many other types of conjoint analysis.

They’ll see a series of sets, showing several concepts at a time. Each concept will have different attributes or levels.

Typically, CBC is thought of as the style of conjoint that most closely replicates a purchase decision process. It asks respondents to make a choice, or no choice, by making trade-offs. There may be a product that has everything they want, but it’s likely the price will be higher.

CBC provides robust results for pricing work, in contrast to adaptive conjoint which we’ll discuss shortly. It is also known as discrete-choice analysis, technically, when it includes the continuous variables of time or price. 

It’s a very popular trade-off technique for new product development research. It will provide reliable data for the relative importance of attributes or levels, based on subconscious rather than claimed thinking. 

Traditionally, CBC uses a full-profile conjoint style, which risks respondent fatigue through exposure to every possible concept combination of attributes and levels. However, CBC is often run using an ‘experimental’ design. This means that only a fraction of the configurations need to be tested, to achieve similar results compared to if respondents saw every possible permutation.

#3 Full-profile conjoint 

To run a conjoint analysis, the online survey is designed to display the full service or product descriptions. By rating or ranking all of the different configurations, the results will statistically show respondents’ preferences.

In a full-profile conjoint analysis, every variable factor in every configuration is presented to respondents, so it’s as comprehensive as possible. 

However, this means that unless you cap the number of permutations to test to a manageable number, a full-profile conjoint exercise will cause respondent fatigue, which will impact response rates and potentially the quality of the data.

Therefore, a full-profile conjoint analysis is better suited when there are only a few factors to test.

#4 Partial-profile / adaptive conjoint types 

These conjoint techniques are better for accommodating higher numbers of attribute levels in a way that is less time-consuming or confusing for respondents to evaluate. 

However, they do this by omitting some of the possible combinations that a respondent will see:

• Partial-profile conjoint: No feature is shown more frequently than any other, but unlike full-profile conjoint, which shows all features in every combination, this technique only shows a select few factors and not every respondent will see every combination.  
• Adaptive conjoint: This uses paired comparisons of partial profiles – it varies the feature sets shown to respondents based on their preferences. However, it tends to underestimate the importance of price as a factor and the technique is rarely used.
• Adaptive choice-based conjoint: This is more reliable for pricing as it uses some choice-based questioning. However, this technique also relies on a certain amount of self-explicated scaling.

#5 Self-explicated scales – quantitative or qualitative

Whereas the above methods are more indirect ways to find out what matters most to your customers, this method asks them directly. 

It’s possible to use fairly standard questionnaires for this approach and it also involves less statistical analysis.

It asks about individual features, rather than bundles of features. Respondents eliminate any that they would not consider in any circumstance, pick the attribute levels they prefer most and least, then everything else is rated relative to those on a scale.

The less statistical approach overall means that it leans more towards analyzing reported rather than subconscious trade-off behavior. Also, as with adaptive conjoint, it’s often ill-suited to analyze the impact of price, because respondents tend to prefer the lower-cost options in this exercise. 

While this exercise is often conducted quantitatively, as per the other approaches, there’s an opportunity to get different kinds of insights – more exploratory and explanatory – through qualitative research techniques. It’s perfectly possible to ask about individual features and make comparisons on a scale during depth interviews, as long as you don’t ask about too many. 

In addition, with qualitative research, the format gives you the time and opportunity to prompt decision-makers about their trade-off choices. This way, you’ll get a better understanding of the reasons why some features are so popular, why some aren’t, and why certain trade-offs are made.

Best practices for trade-off analysis in market research

#1 Don’t include too many features or feature levels

The more you include, the less flexibility there is on the trade-off analysis type you can use. 

In addition, the more features you test, the longer the trade-off exercise is because there will be more sets of questions. That will risk increasing respondent fatigue and decreasing the number of complete responses – either lengthening the fieldwork process, reducing the robustness of the research, or both.

Stakeholder management plays an important role here. A long list of ideas from different teams will need to be scrutinized to some extent, to make sure only the most relevant ones make it through. Presenting respondents with clearer propositions will lead to better results and more actionable insights.

Consider qualitative research at the outset, before using trade-off analysis, to whittle down a long list of features, or to identify better features that directly address customers’ pain points.

#2 Don’t prioritize methodological detail over the big picture

Trade-off techniques are complex and it’s easy to enter into a debate about small tweaks that can be made to alter the methodology, approach, or analysis.

More often than not, minor changes have a minor impact on the overall results. However, they can significantly slow the overall process down and create confusion, particularly if they lead to a lack of consistency.

During trade-off analysis decision-making, prioritize working towards the end goal of getting a robust, statistical market simulation of different factors’ impact on purchases – rather than the precise methodological detail.

#3 Use customer-friendly language to improve accuracy

The accuracy of the results will be compromised if your respondents don’t understand what you’re asking them to make a trade-off decision between – so, avoid jargon.

They’ll likely deprioritize any features they don’t comprehend, but with a better understanding those features could go into their top three – and you’ll never know. 

You could also use qualitative research before starting a trade-off analysis, to test that the language is customer-friendly and how they would be presented in marcomms.

None of this means that you should make the survey too easy for the wrong respondents to take part in though. Make sure that there are still a few ‘red herring’ answer options in the survey and that the screening questions are precise, testing for industry-specific knowledge if relevant.

If you’re running the research in several markets and languages, accurate translations that have also been sense-checked are essential too. Otherwise, the results could suggest that some features are less important in some countries when the reality is that they were just badly translated.

#4 Test the survey thoroughly – make sure the algorithms work

Standard surveys are relatively straightforward to test, as long as you check that the various routes through each survey all work as intended.

However, surveys including a trade-off analysis exercise have more potential to go wrong if you don’t test them thoroughly. They use an advanced algorithm that must show the right features to the right respondents, otherwise, it won’t work and you won’t get accurate results.

Always test to make sure the trade-off technique is working at the front end, as well as at the back end where the analysis will be conducted.