Progressive User Adoption
Published: March 23, 2009
User assistance can add value to a product or Web service’s business model by influencing how deeply users adopt new features or services. As more products employ pay-as-you-go models like that of SaaS (Software as a Service), the contribution user assistance makes becomes increasingly more important.
Users of technology products—from mobile phones to ecommerce Web sites—often stop learning and adopting features long before they’ve mastered those products’ full capabilities. A learning plateau usually occurs once a user has learned the features that meet his minimum product-adoption criteria, when the benefit of adopting more features doesn’t seem worth the extra effort or risk.
A bank’s online bill-paying service provides an example: It is common to find many users paying bills manually online rather than using more advanced features that would let them receive bills electronically or make automatic payments online.
An ecommerce site presents another example: A user might go to the site to buy books, but does not buy other types of products the site offers, such as music or electronic gadgets.
In cases where user adoption curves flatten out at suboptimal levels, companies miss out on revenues they might otherwise get from additional fees or sales. Even in non-consumer-facing applications, suboptimal adoption levels can lead to their economic harm. For example, administrators of a network security system might find the application’s reporting capability to be unsatisfactory, because they use its canned reports instead of learning how to customize reports and improve their effectiveness. Their dissatisfaction with reporting could make the security vendor more vulnerable to competitors who boast of their more useful reports.
Many technology companies think of user adoption as an all-or-nothing, take-it-or-leave-it decision. This column poses an alternative view: progressive user adoption. A progressive user adoption strategy consciously exposes new features of a product and moves users to new levels of product adoption over time. In this column, I focus on the role of user assistance in promoting progressive adoption.
Causes of Suboptimal Adoption
When considering how to persuade users to expand their adoption of more advanced product features, it helps to know why users stop learning in the first place. Based on observations of users during usability tests, in training environments, and at work, my conclusion is that users stop learning for the following reasons:
- Users shift from a learning or exploration mode to a task execution mode. In their initial encounters with a new product, users explore and experiment as part of their learning process. Once users learn enough to meet their initial goals, they stop exploring and experimenting. Instead, they focus on doing the tasks that initially motivated them to use the application.
- The benefit versus effort ratio gets smaller. This ratio becomes less attractive as users move from initial adoption to incremental improvement of their skills. For example, when first confronted with a new phone, users are highly motivated to learn how to make and receive calls and consistently reach those goals. However, the effort to change ring tones, enter contacts’ names and numbers in the directory, or learn how to conference in a third party might be more than a user is willing to make.
Understanding these root causes is one thing; knowing how to overcome them is another. For that, we need a better understanding of the dynamics of technology adoption.
Technology Adoption Models
Two well-established models let us understand user adoption of technology products:
- the technology acceptance model
- the innovation diffusion model
Technology Acceptance Model
Figure 1 shows a slightly modified technology acceptance model that is based on the one Davis, Bagozzi, & Warshaw originally proposed. I’ve added a box labeled Perceived Security, reflecting what later researchers have learned about users’ willingness to conduct ecommerce over the Internet. In looking at this technology acceptance model, one thing becomes clear: To persuade a user to adopt an additional feature, you must affect the user’s perception of the feature’s usefulness and ease of use, as well as the user’s sense of security concerning that feature. In the case of progressive user adoption, user assistance can be an effective external variable that accomplishes all of this.
Figure 1—Technology acceptance model
Everett Rogers identified the following attributes of innovations and their positive (+) or negative (-) influences on adoption:
- relative advantage (+)—The degree to which users perceive an innovation as better than the idea it supersedes.
- compatibility (+)—The degree to which users perceive an innovation as consistent with the existing values, past experiences, and needs of potential adopters.
- complexity (-)—The degree to which users perceive an innovation as relatively difficult to understand and use.
- observability (+)—The degree to which the results of an innovation are visible to others—how easily users can describe them.
- trialability (+)—The degree to which users can experiment with an innovation on a limited basis.
In short, the process for successfully promoting an innovation is as follows:
- Tell how it is better than what a user is doing now.
- Demonstrate that it is easy and consistent with what the user already knows or already does.
- Let the user try it in safe, verifiable increments.
I had an opportunity to see the third point in action, while observing a series of focus groups. The focus groups saw two versions of a user interface. One was very simple, but lacked a robust set of features, and the other offered a robust set of features that market research had indicated users wanted.
The facilitator demonstrated both user interfaces, then asked which one the members of the group preferred. All groups selected the simple one, adamantly claiming the other was too busy. But when asked what they would change about the user interface they’d preferred, they incrementally added functionality that eventually recreated the options they had initially rejected. Then, when shown the rejected user interface again, they enthusiastically endorsed it. Products that look overwhelming and busy, at first, often end up matching the level of functionality users ultimately want. They just need to get to that level in manageable steps—precisely the strategy of progressive user adoption.
A Coherent Strategy for Progressive Adoption
Now, I’ll describe a concrete series of steps you can take to increase user adoption of a product’s features.
First, accept that user adoption is not a single event or decision on the part of a user. User adoption happens in phases, which are affected by the frequency of product use and also determine how many features or services a user eventually uses. A progressive user adoption strategy consciously moves a user to new levels of product acceptance over time, through an orchestrated sequence of exposures to the product’s functionality. A well-formulated progressive user adoption strategy does the following:
- identifies core functionality
- makes the core functionality bulletproof from a usability perspective
- identifies progressive adoption sequences that go from core functions to advanced functions
- constructs product interventions to move users to advanced functionality
Step One: Identify Core Functionality
Core functionality is the basic functionality that, if not achieved, will guarantee the user will reject the product. Understanding this functionality lets you do two things:
- predict where the initial suboptimal learning plateau will occur
- structure the initial user experience around that core to guarantee success, build trust, and establish a functional base on which to add other features
Step Two: Make the Core Functionality Bulletproof
From a product design perspective, initially, the overriding goal should be to optimize the user experience that touches the core functionality. The unspoken corollary here is this: “Don’t break the core functionality as you add features.” Promoting or adding advanced features can also add real or perceived complexity, disrupt compatibility with users’ established routines, and increase a sense of risk. All three of these consequences can inhibit adoption.
Step Three: Identify Progressive Adoption Sequences
Identify the next layers of features and functionality that could represent logical steps for users to take over time, once they have increased familiarity with the product. Here is an example of a revenue-driven strategy an online bill-paying service implemented:
- After signing up for the service, users would typically continue getting their bills in the mail, but manually pay them online. This was the core functionality. If this was not a successful, trustworthy transaction, further adoption would not occur. In fact, users would reject the service outright.
- The bill-paying service would then persuade users to receive their bills electronically, but still manually pay them online. This electronic billing feature offered important revenue opportunities for the bill-paying service, allowing it to extend its revenue model by charging billers for the convenience of not providing paper bills. Early attempts to get users to sign up for electronic billing during their initial enrollment had not been very successful. The bill-paying service found that users needed to first develop trust for the online bill-paying process in general.
- Next, users who had authorized their receiving bills electronically would further authorize the automatic payment of those bills online. Again, the adoption of this feature had positive revenue implications for the bill-paying service, ensuring that every payment to a biller would go through its online service—and thus, incur a transaction fee.
Step Four: Construct Product Interventions to Move Users to the Advanced Functionality
An important goal of this step is to identify moments of opportunity that indicate readiness on a user’s part to advance to a new level of functionality. The complexity this step entails can range from simple front-end logic, which executes in the user interface, to complex back-end processing and decision-making engines.
For example, in the bill-payment application I described earlier, a significant moment of opportunity for automated payments would occur if a user made the same payment online to the same biller for several months in a row. The application could easily identify that pattern from payment history data it already presents in the user interface. At the more complex end of the spectrum, consider the sophisticated data processing Amazon.com uses to make recommendations, according to both the buying patterns of the current user and the aggregated patterns of its larger user base.
Interventions can take many forms—from dialog boxes to dynamically constructed screens that provide the appropriate messaging and options for the opportunity at hand. What is of tantamount importance is that the intervention not be overly intrusive, impacting the core functionality. As much as it might go against the grain of the marketer on the product team, you must make it easy for a user to ignore the intervention. By all means, make it noticeable. Just don’t force users to change their habitual task flows in order to reject the option or suggestion. Beware of the Clippie effect, in which well-intentioned user assistance interventions became aggravators rather than sources of delight.
Figure 2 shows an example of an intervention in which biller information appears if a user moves the focus to a bill-payment group, prompting the user to use an advanced bill-paying feature—electronic bill presentment. The Biller Information box gives the user information about the biller and the past transaction history, increasing the user’s confidence and trust in the system as a whole. This increased trust reduces perceived risk in adopting electronic bill presentment.
Figure 2—Example of a progressive adoption intervention
Figure 3 shows a tip that encourages a user to use the advanced functionality and an associated link that makes it easy to do so. The application displays these elements dynamically, based on the selected biller and a set of predefined adoption sequences.
Figure 3—A helpful tip
Usability testing of an actual design that was similar to that shown here ensured that the intervention’s presence did not interrupt the habitual bill-payment task flow, which occurs on the left-hand portion of the screen. The successful design of progressive user adoption relies on usability regression testing to ensure the user experience of the core functionality does not get broken when adding interventions to encourage progressive adoption.
A strategy of progressive user adoption lets a company leverage its established user base to increase revenues and solidify its competitive advantage, without requiring expensive development of new features or costly user-acquisition campaigns. The primary application of progressive user adoption interventions is in usage-based revenue models, but license-based models are also a target in cases where increased adoption secures a more loyal customer base.
The overall strategy for progressive user adoption starts with the solid foundation of a satisfying user experience of a product’s core functionality, then builds a logical progression from that base by identifying moments of opportunity and appropriate interventions. Well-crafted on-screen messages and user assistance are critical components of effective interventions and need to take into account both the technology acceptance model and the attributes of innovation diffusion I’ve discussed in this column. Not disrupting the core user experience is a key requirement for those interventions.
Davis, Fred, Richard Bagozzi, and Paul Warshaw. “User Acceptance of Computer Technology: A Comparison of Two Theoretical Models.” Management Science, August 1989.
Nilsson, Maria, Anne Adams, and Simon Herd. “Building Security and Trust in Online Banking.” Conference on Human Factors in Computing Systems, CHI '05 Extended Abstracts on Human Factors in Computing Systems, Portland, Oregon, April 2005.
Rogers, Everett. Diffusion of Innovations. 4th Edition. Glencoe, Illinois: Free Press, 1995.
Salisbury, W. David, Rodney Pearson, Allison Pearson, and David Miller. “Perceived Security and World Wide Web Purchase Intention.” Industrial Management & Data Systems, Volume 101, Number 4, 2001.