Definition: Conjoint is a survey-based research method that helps managers understand what product features and price points drive customers’ decisions to buy a product. It was developed in the 1970s and still remains one of the most widely used and robust techniques in market research. The basis of conjoint is breaking down products into their characteristics (such as price, shape, colour, features) and then trying to estimate how important each of these characteristics is using a carefully designed survey that aims to simulate respondents’ choices. Common applications of conjoint are pricing, product feature selection, and advertising claim assessment. Learn more about conjoint analysis.
Most marketers will have heard of conjoint analysis, many will have commissioned a study from a research company, and some will have done it themselves. The one thing that probably strikes people (apart from the power of the method in eliciting drivers of customers’ choices) is how long and expensive it is. It is not uncommon for researchers to take two to three months to run a project for tens and in some cases hundreds, of thousands of dollars. In today’s world of quick insights and rich readily-available data, we tend to put conjoint into the too-tough bucket reserved for large strategic pricing projects.
While this perception is lasting and pervasive, a few companies are finding that conjoint (and other advanced analytics methods) can be done faster, cheaper, and more often, helping them understand customers’ behaviour on a new level in a much more agile way than before.
Here, we will go through what conjoint is (and why it is so powerful), traditional applications of conjoint, and some exciting new ways to use the method.
What is conjoint analysis again?
Conjoint analysis is a market research method that views products as combinations of product attributes. For example, a team chat app, such as Hipchat or Slack, can be represented as the following combination of attributes:
- App name
- Maximum number of group members
- Time of retention of messages
- Searchable history
- File sharing capability
- Monthly price
These attributes are essentially what users see on product comparison websites or the pricing pages of SaaS companies. Each of these attributes has multiple “levels”: e.g., the monthly price can be “Free”, “$2 per team member”, “$9 per team member”, or “1,000 per user group”.
The main idea behind conjoint is that each of these levels has intrinsic value (or “utility”) to the buyer and the utility of a product is the sum of the utilities of the levels that make up this product. In other words, the higher the value of the individual features that make up a product, the more likely a customer is to choose that product.
In order to estimate the utilities of the product features, we ask (typically about a hundred) respondents to answer roughly a dozen questions. Each of these questions presents the respondent with a choice from a set of three to five products. Each of these products will be a combination of the various levels. In other words, respondents are asked to CONsider JOINTly the various product features and competitive brands, just like they would in a real life, outside the survey environment.
Then magic happens. A Hierarchical Bayesian Model “uncovers” the utilities of each level to each respondent based on their responses to conjoint questions. This analysis produces a number of useful outputs:
- Importance scores of attributes;
- Utility scores for each level;
- An interactive market share simulator, which is helpful in a number of ways:
- you can play with your product features and choose the ones that will maximize your own market share,
- if you are doing pricing, it can help you find not only market-share-optimising price level (which is likely to be the lowest price), but also help you find a profit-optimising price point,
- if you are considering possible competitors’ moves, you can estimate how much your own share they can eat and what changes to your product you need to make to fight back;
- Willingness to pay for a feature;
- Price elasticity of demand;
- Segmentation of customers based on preference for product features.
Conjoint is not the only way to understand what drives your customers’ decisions, segment them, and predict market shares, but it is most widely used and well supported by decades of marketing science. Without going into theoretical depths, it is fair to say that fewer management ideas have such strong following both in business and academia.
How is it used in 2017?
Pricing is the most typical application of conjoint analysis, particularly in FMCG, financial services, and telecommunications. A typical conjoint project would be run over the course of two to three months at a fairly high cost (starting at $20,000 and going into hundreds of thousands). This works well if your business is in a slow-moving product category, your competitors are operating at a similar pace, and no decision in your organisation takes less than three months.
Trouble is such sanctuary environments are becoming extinct. Marketing is becoming faster and automated, competition is fierce and comes from unforeseen angles, and as a result, consumers’ preferences and their decision-making are fickle. These days, even large FMCG companies start to realise that a conjoint (or any other big market research project) that is run today is a poor predictor of customers’ preferences in 12 months’ time, when the factory is finally set up to manufacture the new product, but competition has gone ahead and taken consumers with it.
The answer, of course, is to get nimble, agile, and smart by running smaller and timelier conjoint studies, which is now possible with software products that are currently on the market.
How will conjoint be used going forward?
Even though conjoint is still primarily used to define features and prices of products in advance of their launch, or in preparation for a major overhaul of a product line, new ways of using the methodology are emerging and gaining traction.
Iterative feature selection within Agile product development. Few modern marketers refute the ideas of Agile product development, one of the key principles of which is listening to your customers’ feedback and re-prioritising your development plan accordingly. Yet, how do you know which feedback to listen to? Should you spend two sprints on improving feature A or one sprint on building feature B?
In most cases, these solutions are clear, but once in a while even the most intuitive product manager or marketer will need to probe their customers a little more to understand what features they are prepared to pay for or what will sway them away from competition. Asking customers what they prefer in a simple survey does not always work because people do not fully know their own preferences. Luckily, conjoint is here to help product managers who need to quickly gather insight to plan their sprints.
Frequent pricing optimization. In some fast-moving industries, such as electronics and software, companies are used to quick competitive moves and customers expect to get new product releases every few months. But few industries see faster rates of new product releases than gaming. That is why gaming marketers are discovering conjoint analysis as a way to confirm (or disprove) their hypotheses on the right pricing and see how sensitive the market is to competitors’ new launches. In this industry in particular, marketers are faced with changing preferences that require almost monthly choice modeling exercises.
Brand performance measurement. Brands’ strength is flighty in the hyper-connected world. With just one controversial or powerful ad, a company can jeopardize or elevate its brand’s perception, fuelled by the raging and raving of social media. While there is no shortage of tools for measurement of brand strength, few of them can easily translate findings into dollar terms. Conjoint’s ability to measure the elasticity of demand comes handy for advertising tests and branding studies, for the higher the brand the lower is the elasticity of demand for it.
These new applications of conjoint analysis are only possible today because of automation of the mechanics of setting up a study (a very hard thing, if you ask any market researcher), data collection, and analytics. Thanks to services like Conjoint.ly, what used to take months, now takes days. What required a qualified statistician, is now available to any marketer. With the new tools at your disposal, how can you use insights from the conjoint analysis in your marketing?