At Usability Sciences we try to help our clients understand, good research projects should not only provide answers to your research questions, but should also lead to the generation of new research questions. Research always begets more research. In academic circles, they refer to this as “heuristic provocativeness”. A project is not an end it is the beginning of a journey. The actionable findings from one study should answer the research question(s) and offer avenues for future inquiry.
Tackling the Persuasion Puzzle
The primary goal of most businesses is to sell products or services as effectively and efficiently as possible. It is a game of persuasion and persuasion is a science. We view CX/UX research as a piece of the persuasion puzzle contributing to the goal(s) of the business. This is a constant process; no organization ever finalizes their persuasive model. Until you have an algorithm that predicts human behavior with 100% accuracy* and can apply the algorithm with 100% accuracy, there is always room for improvement.
*To put the 100% accuracy goal in perspective, some of the largest effect sizesin the social sciences are in the r2 = .30-.40 range. An r-squared of .40 means the model is explaining 40% of the variance in the sample on the dependent variable(s). Meaning, with the variables we accounted for, our algorithm can predict 40% of the difference between time one and time two, or groups (if we are using eta squared). These are models with decades of academic inquiry from multiple research teams examining countless contexts and additional explanatory variables. Models that scholars have dedicated entire academic careers to. Models that address deeply rooted and powerful psychological constructs, such as the fear of death, and they still only explain 30-40% of the variance, that is, are 30-40% accurate. While something like explaining 40% of customer behavior may not seem worthwhile, consider how much of your customers’ behavior you are likely currently explaining and what something like a 3% total increase in predictive power would mean for business revenue.
To gain predictive power, CX/UX teams need to avoid tunnel vision, by considering how their research contributes to a holistic persuasive model for the business. This can be a challenge if you are primarily thinking of your projects in isolation, as “one-offs”, rather than looking to make continual incremental increases through sustained research programs. For instance, when you conduct a usability validation test, if you only focus on users responding positively to your system, favorable SUS and Net Promoter scores, you are missing a wealth of information and insights. These projects are not just means to an end, they are learning opportunities that provide you with evidence for how users behave in the context of your business. The implications of these learnings stretch far beyond the singular research exercise to validate your design. Each project is another piece of the puzzle learned and contributes to a body of research specific to your organization or product, like a line of academic inquiry.
Building a Body of Evidence
The research you do today becomes data for tomorrow, to be utilized in guiding design and generating research questions and hypotheses for future research. If you invest in quality research, you should be confident in referring to the report(s) to rediscover information about how your actual users behave in general.
It is crucial the research conducted is rigorous and utilizes the appropriate methodology to ensure these findings provide foundational insights. A proper research project will attempt to identify the context in which the elements in a system worked or failed, not just the user response to individual elements. A menu icon in one context may not work (e.g., is not findable, is not intuitive), while the same menu icon in another context provides the optimal experience. Good research should help you understand the interactions of elements and context such as, X element given Y and Z conditions is effective with your users and not effective when condition Z is absent.
For example, consider that your research found a slider on a before and after image (X) works for market segment (Y) when an animation showing how to use the slider is included in the design (Z). It’s not sufficient enough to say, “Ok, we’ve validated the animation. Ship it.” This finding should also lead to you asking more questions such as, “If we include a similar component elsewhere, how do we incorporate animations to effectively teach users to navigate the system?” Or, perhaps, “An animation is required to help segment (Y) succeed, but does it impact the credibility of the before and after images for the other segments (A,B,C) that have a higher lifetime value?” And so forth.
These findings help you establish “best practices” for your business context (i.e., your customers, industry, and sales propositions). When you approach research this way you begin to build a body of evidence and increase your predictive power to persuade customers to engage in desirable behaviors.
Monitoring the Ever-evolving Marketplace
However, the value and importance of sustained research are not due solely to the ability to ask follow-up questions and drive additional depth. Regular research is required due to the rapid pace the market moves. Your customers, industry, and sales propositions are constantly changing. The foundational insights you uncovered from your last research project are not guarantees, they are guides to be used by a skilled CX/UX team. You must test and re-test findings continuously to maintain your predictive power. Replication of findings is crucial and is something often overlooked in the field of CX/UX research. One study that provides a foundational finding is good, a second study corroborating the finding is great. This does not mean you need to repeat the same study twice. It means, past key findings should be included in future research projects as additional areas to cover when possible.
Avoiding Cross-functional Pitfalls
Finally, CX/UX teams need to understand and find ways for their research efforts contribute to a larger business intelligence team. One way CX/UX professionals can contribute to a holistic predictive model and avoid the tunnel-vision trap is to ensure their research efforts coordinate across functions and teams. However, many organizations silo various research efforts and only pay lip-service to the idea of “cross-functional” teams. The care teams get direct feedback from customers, the web-analytics team track descriptive statistics, the data scientists perform some inferential statistics with DOE analysis, the customer insights team often focuses on market research, and the UX team is the empathetic voice of the customer often primarily doing applied qualitative research on systems.
Creating cross-functional teams and having all or most of these departments roll-up into one executive is not efficient in moving the dial (i.e., increasing your predictive power) unless the specialized experts on the team understand the business’ goals, how their colleagues contribute to a holistic persuasive model, and how to leverage each approach. If the data scientists find conversion increases given X and Y conditions despite users insisting they do not like X and Y conditions, the CX/UX team needs to take these findings and uncover what is really happening and work with the designers to create a solution that optimizes conversion and user experience before the competitor does. Further, your CX/UX insights can be tested and replicated by tying elements to KPIs and having the data scientist perform DOE analyses while the customer care team continuously monitors negative feedback to the changes.
* * *
At Usability Sciences we not only provide world-class CX and UX research to answer your research questions and test your hypotheses, we understand how these various departments intertwine and can equip our clients with insights to guide next steps to increase their holistic predictive model. We are multi-lingual specializing in both qualitative and quantitative analysis, and we would love to be your business intelligence partner.