SurveyMonkey her tür senaryo ve ihtiyaç için tasarlandı. Nasıl yararlanabileceğinizi öğrenmek için ürünümüzü keşfedin.

Anket alanındaki dünya liderinden veriye dayalı görüşler elde edin.

100'den fazla uygulama ve eklentiyle entegre edin.

Bilgi toplamak ve ödeme almak için özel çevrimiçi formlar oluşturun.

Yerleşik yapay zeka ile daha iyi anketler yapın.

Tüm pazar araştırması ihtiyaçlarınız için özel çözümler.


İşletmeniz için müşteri memnuniyetini ve sadakatini ölçün.

Müşterileri mutlu edip marka savunucusu yapmanın yollarını keşfedin.

Daha iyi kullanıcı deneyimi için uygulanabilir görüşler alın.

Müşteri adaylarının ve davetlilerin iletişim bilgilerini toplayın.

Sıradaki etkinlik için kolayca LCV'ler toplayın ve bunları takip edin.

Sıradaki etkinliğin daha iyi olması için katılımcıların isteklerini öğrenin.

Bağlılığı artırmanızı sağlayacak görüşleri açığa çıkarın.

Daha iyi toplantılar yapmak için katılımcılardan geri bildirim alın.

Performansı artırmak için iş arkadaşı görüşlerini kullanın.

Daha iyi kurslar ve eğitim yöntemleri oluşturun.

Öğrencilerin ders materyallerini nasıl değerlendirdiğini öğrenin.

Müşterilerin yeni ürün fikirleriniz hakkında ne düşündüğünü öğrenin.


Anketleri ve verileri kullanmak için en iyi uygulamalar

Anketler, iş ipuçları ve daha fazlasıyla ilgili blogumuz

SurveyMonkey'i kullanmak için eğitimler ve kılavuzlar

Büyük markalar SurveyMonkey ile nasıl büyüyor?

Satış Ekibine UlaşınOturum aç
Satış Ekibine UlaşınOturum aç

How to prioritize features using MaxDiff analysis

What features do your customers value? Find out using MaxDiff analysis from SurveyMonkey.  

No market research for a product is complete without MaxDiff analysis. Using a unique type of survey question, you can delve into what your customers consider to be the most and least important attributes of your product. MaxDiff analysis will help you quantify your target market’s preferences, so you can develop the most important features first.

Let’s take a look at MaxDiff methods and how you can use them to prioritize your product features.

MaxDiff (maximum difference) analysis, also known as Best-Worst Scaling, is a trade-off research technique used to quantify preferences. It helps you to sort features, messaging, brand names, and more, according to customer preferences. MaxDiff also allows you to test combinations of similar features. By identifying what consumers want most, you can maximize your time and capital on efforts that truly affect your customers’ purchasing decisions.

Keep in mind that although MaxDiff provides quantitative data, your results can also be considered qualitative because the analysis shows how important your attributes are relative to each other, not how important they are in general.

MaxDiff analysis uses survey research methods to ask participants to choose among several options at once by selecting only the best and worst options. By eliminating potentially vague mid-range responses, there’s no guesswork involved in understanding what participant answers mean. 

MaxDiff analysis may bring conjoint analysis to mind. The methodology in conjoint analysis is similar, but instead of going through several iterations of tradeoffs, MaxDiff limits the number of decisions the participant has to make but still provides actionable insights.

Best-worst scaling questions use attributes (a property, item, or feature to be measured) and sets (a group of attributes provided to participants). MaxDiff questions can be asked as part of a longer market research questionnaire or in a standalone survey. 

Traditional ranking and rating questions, such as Likert scales, allow survey respondents to choose along a scale or rank items on a continuum. MaxDiff forces respondents to choose only the most and least important options. This type of analysis is immune to user scale bias, scale meaning bias, lack of discrimination, and other types of survey bias that can plague rating and ranking questions.

Example of a traditional ranking question

Please read through the following list of smart television features and rank them in terms of your priority.

4K UHD streaming1
Music streaming3
App store5
Voice control4

Example MaxDiff question

Please indicate the smart TV feature that would be the most important in your purchase decision and which feature would be the least important.

Most ImportantFeatureLeast Important
◻️4K UHD streaming◻️
◻️Music streaming◻️
◻️App store◻️
◻️Voice control◻️

MaxDiff is critical for finding out what your target customers value most. There are several advantages to using MaxDiff.

  • Easy to answer: Respondents will have no problems answering the best-worst questions. The decision-making process in MaxDiff is similar to real-life situations that require people to make choices and trade-offs as they eliminate options. Respondents’ answers to a series of these questions reveal the strength and importance of each choice.
  • Reduces bias: MaxDiff questions avoid cultural, list order, scaling, and other biases that can affect the results of rating and ranking questions. By reducing the risk of bias, businesses won’t waste resources on features that aren’t the ones their target market wants.
  • Increased customer satisfaction: When you take action based on MaxDiff survey results, your customers will experience increased satisfaction with your product and brand. 
  • Test more items: It’s simply easier for respondents to choose the best and worst options rather than ranking, for example, 10 attributes. Ranking questions with several attributes can lead to survey fatigue and low-quality data. 
  • Quantify results: MaxDiff survey data is suitable for creating statistical models. The models allow you to quantify the preferences of your target customers.

Now that we’ve established what MaxDiff is and the advantages of using it, let’s move on to learning about feature prioritization with MaxDiff analysis. While you can do this on your own, a survey tool, like the Feature Importance MaxDiff tool from Momentive, maker of SurveyMonkey, is useful for analysis and reduces the risk of human error.

  1. Decide which features to be included in the MaxDiff analysis

Before you can begin to develop your survey questions, you need to determine which features you want to evaluate. Sit down with project stakeholders and choose 12-25 features. This range offers the best number of comparisons to determine each feature’s relative importance.

  1. Create a survey and collect responses

You can design your own study or create it with the assistance of the market research consultants at Momentive. Your survey should offer attributes divided into subsets, with one subset of three to five features provided for each question. Using a tool helps you to generate subsets randomly. This allows you to obtain the most accurate, actionable data. 

If you’re creating your MaxDiff analysis manually, keep in mind that the experimental design must contain:

  • Frequency balance: each attribute appears as a choice an equal number of times
  • Orthogonality: each attribute is paired with others an equal number of times
  • Position balance: each attribute appears the same number of times in each position

Use the following formula to determine how many sets are required:

# of times to show attribute = total attributes / # of attributes per set

Following these guidelines means that respondents will see each feature once during the survey, but not necessarily every combination of features in a set.

  1. Analyze the results

If you use a tool, it will provide you with a statistical analysis of your data. For manual analysis, you’ll need to calculate the MaxDiff. 

  • Create a table to record your responses. Your table should look something like this:
Total Responses200
FeaturesMost ImportantLeast ImportantMagnitude
Feature 110020
Feature 24020
Feature 32060
Feature 440100

The data in this table indicates that 100 respondents said that Feature 1 is the most important feature and only 20 of 200 selected it as least important.

  • Next, you can calculate the relative preference magnitude of each feature to interpret how the features should be prioritized. Use this formula to calculate relative preference magnitude:

(most important - least important) / total responses

Your table should now look like this:

Total Responses200
FeaturesMost ImportantLeast ImportantMagnitude
Feature 1100200.4
Feature 240200.1
Feature 32060-0.2
Feature 440100-0.3

Magnitude is measured on a scale from -1 to 1. The closer the magnitude is to 1, the stronger users prefer that feature. Data from this calculation can be put into a graph for easy visualization.

  • Alternatively, you can find the utility (or value) of each feature using MaxDiff scores by subtracting the number of times an answer was chosen as least important from the number of times it was chosen as most important. Arrange the values around a mean to determine relative value.
  1. Put the findings into play

Once your analysis reveals which features your customer would value most, take your data and analysis back to stakeholders. You’ll be able to make informed decisions about prioritizing your feature selections.

Let’s look at examples of MaxDiff determining feature prioritization. 

Example 1

You are in the hospitality industry and are in charge of remodeling one of your resorts in the northern US. It’s a costly venture, so you want to know what is most important to vacationers in your target market.

You create a list of features you might want to upgrade in the resort:

  • Indoor pool
  • Workout room
  • Business center
  • Room service
  • Weekend deals 
  • All-inclusive options
  • Free breakfast
  • Early check-in
  • Outdoor pool

Question 1

Of the following features, please indicate which is the most important and which is the least important when choosing a vacation resort:

Most ImportantFeatureLeast Important
◻️Indoor pool◻️
◻️Workout room◻️
◻️Business center◻️
◻️Room service◻️

Of the following resort features, please indicate which is the most important and which is the least important when choosing:

Most ImportantFeatureLeast Important
◻️Weekend deals ◻️
◻️Free breakfast◻️
◻️Early check-in◻️
◻️Outdoor pool◻️

The remaining questions will continue with various subsets comprised of the features you want to evaluate. The questions will include frequency balance, orthogonality, and position balance.

You submit your MaxDiff survey to 300 members of your target market. The data reveals that, of the 300 responses, most participants consider an indoor pool and a free breakfast to be the most important options. 

This data is presented to management, and feature prioritization is decided based on the data.

Example 2

The airline you work for wants to know what features will make customers choose them over the competition.

You create a list of features that could be initiated or changed at your airline:

  • More legroom than competitors
  • Free in-flight meal
  • Mileage program 
  • In-flight entertainment
  • Free checked bag
  • Free carry-on bag

Question 1

Which of the following features would be most likely to factor in your choice of a particular airline, and which would be least likely to make you choose it?

Most ImportantFeatureLeast Important
◻️More legroom than competitors◻️
◻️Free carry-on bag◻️
◻️Mileage program for earning miles◻️

Question 2

Which of the following features would be most likely to factor in your choice of a particular airline, and which would be least likely to make you choose it?

Most ImportantFeatureLeast Important
◻️More legroom than competitors◻️
◻️In-flight entertainment◻️
◻️Free checked bag◻️

Your remaining questions continue with various subsets comprised of the features you want to evaluate. The questions will include frequency balance, orthogonality, and position balance.

You submit your MaxDiff survey to 500 travelers in your target market. The data reveals that, of the 500 responses, most travelers would prefer more legroom significantly more than any other feature.

This data is presented to management and target flights are chosen to increase legroom.

In addition to feature prioritization, MaxDiff analysis has several other uses. Consider this list:

  • Testing package designs
  • Product name testing
  • Loyalty program offerings
  • Promotions
  • Uncover what messaging and product claims resonate with your target audience
  • Determining tradeoffs customers will make to get features they want
  • Testing ad claims
  • Design concepts
  • Positioning statements
  • Topics of interest

Essentially, any time you want to know your target market’s preferences, a MaxDiff survey will work for you.

Don’t waste time and money on features that aren’t important to your target market. Administer a MaxDiff survey or add MaxDiff questions to your current market research to gain actionable insights into what’s important to your customers.

SurveyMonkey is here to help you with all of your market research. Take a look at our wide variety of market research solutions for everything from branding to sending surveys.

To read more market research resources, visit our Sitemap.