The Art & Science of Expectation Management

The technology and library sciences edition.

December 21, 2015

Let us face it - the expectation from our faculty, student, staff and alumnae in terms of technology and library services is sky high and all of us are struggling to figure out how best to support them.

This is a tough nut to crack. In fact, this kind of complicated multidimensional puzzle could be classified as an NP complete problem. Unfortunately, you cannot learn how to master this by going to school and getting a degree in expectation management (that I am aware of). The only solution is to simply figure it out.

There is, however, little coherence when it comes to meeting the expectations of the stakeholders. Our constituents are diverse in several dimensions - age, gender, technical expertise, appetite for change etc. - and while it might be a natural impulse to try to find a coherent set of expectations about the services we deliver, the sheer diversity makes such a tract a non-starter. Let us look then at the art of involving the community in managing expectations.

Governance. Convening a group of interested representatives from the faculty, staff and students and having them help decide on a viable short and long term plan is an essential part of expectation management. Finding a “group of interested representatives” is much easier said than done and it depends a lot on the institutions. In general, these groups are formed and tend to be very active when there are crises. One needs to figure out a way to have these groups active on a regular basis. Having two groups, one that concentrates heavily on academic matters and another on administrative tend to work the best. The IT/Library leadership should invite members with different viewpoints and not seek out only those who will agree with the leadership most of the time. Creating such structures with the blessings of senior leadership has definite advantages. When a senior leader falls in love with a technology and wants the IT organization to implement it, one can point to the need to run it through the governance groups, provided the campus community has developed trust in the groups and processes.

At Wellesley, we have the Advisory Committee on Library and Technology Policy which consists of six faculty members, five administrative staff and two students. This group meets thrice a semester and we keep the group informed of all major initiatives and policy development and seek their advice. For example, it is this group that helps decide that the College should move to Google Apps for Education approximately five years ago, after examining Google, Microsoft offerings and running Zimbra in house. We also have a group of administrative department managers who help us prioritize the administrative projects that come our way. We are very careful in all these cases not to overload them with unnecessary work or details and are not afraid to cancel meetings if there is nothing to discuss!

Relationship building (and getting out there). Technology and library leadership  should be out there meeting with the faculty, staff and students rather than cooped up in their offices. It also helps if the leaders attend a few classes periodically to learn what kind of role technology plays in the classroom and how well it works (or not). In the same way, attending departmental and divisional meetings to address questions that the administrative staff have or seek out how the services can be improved is also important. As hard as it may be, try to reach out to the students in any way you can. If you have a student government, going to their meeting and hearing their concerns out would be seen in a very positive light.

Of course, in each of these cases, one needs to remember that everyone wants to be heard and everyone feels that their requests will be satisfied. You need to be honest in explaining to them what can and will be addressed and why some of the others cannot be. I have found that most of our users will understand the rationale for why it is difficult to satisfy their requests if it is explained carefully and in the right context. And at some point, you should be bold enough to say “I am sorry that you feel that this is critically important, but I have tried very hard to explain why it is difficult to accomplish. Let us just agree to disagree on this one”. Again, I have found that people appreciate such honesty than trying to give them an impression that their request will be satisfied and in reality it cannot or won’t be.

At Wellesley, my senior team and I offer to visit academic and administrative departments to hear about their needs. Every semester, a few departments invite us. I usually tell them that I am most interested in knowing what things don’t work or how we can improve a service rather than hearing how well we are doing! After every such meeting, we create an action plan that identifies the staff responsible  for each of the request and by when and share with the department head or the chair.  This helps build trust and can create tremendous goodwill. And of course, the onus is on us to make sure that they are completed as promised.

Moving on to the science of it…

Data and analysis. Data can help manage expectations when it is used in conjunction with a couple of things mentioned above. Using valuable survey data such as EDUCAUSE Core Data Service, Campus Computing Survey, Oberlin Group survey and MISO survey can be tremendously important in explaining how your service delivery compares to the appropriate peers. If it exceeds, you should highlight that. If it is below what the others are able to accomplish, you should have a viable set of reasons why and a plan to achieve parity with others. Similarly, collect and use the data to support any of your proposed initiatives. Don’t be fearful of accepting mistakes if the data shows that and plan and execute a course correction.

We have used data in very effective manner. For example, we used LabStats data to determine that the public PCs in our library are an overkill for what the students need. We replaced them with Chromeboxes, thereby providing a much better user experience at a reduced total cost of ownership.

Here again, do not try to use and interpret the data all alone, because there will be plenty of community members who will have probing questions about the validity of your interpretations. There is a general bias to pick and present data that supports our initiatives. I am always reminded of Prairie Home Companion’s claim of “all the women are strong, all the men are good looking, and all the children are above average.”  Partner with the members of the governance groups or other interested faculty (who could potentially use the data for their academic research) to analyze and present the data because it will carry more weight and is likely to be more unbiased.

Accept the fact that most things in life are normally distributed. The exact shape and characteristics of the distribution are very time dependent. For example, if you are in the middle of an LMS or ERP implementation, it may be skewed in one direction than if you just completed a successful implementation. And the tails are fact of life, so accept it! Use weighted tails to manage processes and expectations. What I mean is, those representing the positive tail will rarely compliment you or stand up in a faculty meeting to offer their appreciation, but those on the negative tail are not shy to be heard. Most of the time greasing the squeaky wheel is seen as a solution. At best, it can be a short term band aid, but not a strategy. Developing a sound plan (a normal distribution with a positive skew) and being patient for it to develop should be the strategy while the local fluctuations in the tails are appropriately contained.

Eternal optimization. Finally, I always look at the job in front of us as a pretty complicated constrained optimization problem - we are trying to maximize the satisfaction of the users’ expectations with innumerable constraints some of which may be orthogonal (exactly opposite and therefore cannot be satisfied simultaneously).

The major constraints are typically the resources (both financial and human, as well as an intangible called cultural factors) and the very definition of the “satisfaction”. And this multidimensional surface is constantly fluctuating and even if you have managed to get close to the maximum, there is no guarantee that you will stay there for longer that a fleeting second before the function requires you to solve for the next maximum.

This is why the next batch of CIOs should be required to have a background in operations research!

Ravi Ravishanker is Chief Information Officer at Wellesley College.


Back to Top