That’s K.P.O.A.S.P.P. for those of you playing along at home.

Anyway, here’s how I ended up writing up our issues with academic software purchasing. I think it’s got to be hitting other places as well. More stuff, more complexity, more concerns, and less people and funds can’t be unique to our organization. Even without financial pressure, the support of centralized academic software is going to be an ongoing and expanding challenge. This doesn’t take into account what it would take to support a thriving experimental space.

I have to re-write this in more active voice. I’ve also got some stuff around what we’re going to do about it that I’ll probably post at some point.

Problem Statement

Increased demand for educational technology, complicated by budget and support staff reductions, requires a new process where academic leadership has an increased role in making strategic funding decisions, helping define support levels, and participating in consistent and ongoing evaluation processes.

• Demand: General academic technology use has increased in terms of scale, scope, and complexity. It has become more important to the strategic mission of the institution and the learning goals of individual disciplines. Midd Data and the expansion of online learning at MIIS are two recent initiatives that have significantly expanded support and infrastructure needs.
• Funding: We have been unable to expand the academic software budget to meet requests made over the last four years. The current process does not allow for strategic academic decisions around software additions nor does it provide a process to reallocate existing funds based on institutional needs.
• Support: Faculty expectations around support differ from the support units’ ability to provide those services. There is a need to more clearly establish, communicate, and document technical, training, and pedagogical support levels for academic software. This includes existing software, future purchases, and any supported use of free applications (open source, free/freemium etc.) which fall outside the purchasing process.
• Evaluation: Pre-purchase software evaluation (security, privacy, accessibility, and contract issues) has become increasingly important (legally, ethically, and politically). These reviews require an increasing amount of time and specialized knowledge. Ongoing evaluation of purchased software effectiveness relative to established goals will also be key in determining whether to extend contracts, evaluate new solutions, or redirect funds.

Demand

General academic technology use has increased in terms of scale and scope. It has become more important to the strategic mission of the institution and the learning goals of individual disciplines.

• Use of academic technology increased during COVID and expectations from faculty and students have increased.
• Institutional needs around online learning have resulted in increased needs for online course software and infrastructure.
• An institutional focus on data and data analysis has resulted in increased demands for software and hardware.
• Discipline-specific software options and the desire to use them have increased.
• Increased demand for high performance computing and artificial intelligence will create additional infrastructure and support needs.

Funding

Yearly increases

Even with no new purchases, subscription-purchase models with built-in yearly increases result in increased academic software costs each year. The majority of academic software is licensed in this way.

Below are yearly price changes for a limited selection of academic software.

<<There was a table of prices here indicating that they all go up over time. Those costs add up. The more stuff you have, the greater the cumulative 3 to 5% increase is . . . two plus two is four, minus one that’s three, quick maths.>>

Funding reallocation

If funds beyond the cost of inflation are not regularly added to this budget line, the only other way to provide additional software is to reallocate existing funds by discontinuing support for existing software. There is no established process to evaluate software in this manner. Currently, software does not have specific goals, specific data around those goals, or explicitly defined owners. That makes it very difficult to evaluate existing purchases, make strategic decisions about whether those funds should be used in other ways, or organize a process to reallocate funds from one piece of software to another.

Funding retention

Even if we had a system to make choices about software funding reallocation, we do not currently have a method for keeping those funds within the academic software budget line. In an environment where budgets continue to shrink, it will be important that these funds remain focused on academic software.

Support

Currently, the Help Desk website does not explicitly and publicly communicate support levels for academic software. When new software is added, support expectations (technical, pedagogical, training) are not evaluated against existing support staff workloads. This results in differing support expectations from instructors and unclear workloads for support staff.

Support for software differs markedly by application.

• Increased support staff turnover coupled with increased demands on fewer positions requires more structure, communication, and strategic choices around Middlebury’s internal support processes and patterns.
• It is important to leverage external support services when possible, especially when Middlebury pays for that support (24/7 Canvas support for example).
• Middlebury support staff currently reside in at least four different groups. This disaggregation results in an increased need for coordination and communication.
  • DLINQ
  • ITS
  • Library
  • Sciences Technical Support Services
• The complexity and specificity of discipline-specific software has led to department-level support (GIS, astronomy, translation, film and multimedia culture).

Evaluation

Pre-purchase evaluation

Prior to committing to a purchase it is important to evaluate the software thoroughly with regard to security, privacy, and accessibility. These reviews take time and require a specific set of knowledge and skills which are not widely available at Middlebury.

• The lack of skilled individuals needed to complete reviews results in significant delays in the purchasing process or assessments are completed by individuals without the necessary expertise.
• Software has been purchased without these reviews. That software needs to be reviewed and the results of the review documented.
• As software is regularly changing, regularly repeating the reviews is necessary at defined intervals. As our total software footprint increases, this additional load needs to be considered in addition to the requirements of new purchase analysis.

In addition to these key legal and ethical requirements, it is important to evaluate software capabilities against existing approved solutions. Integrating additional software with similar capabilities results in increased costs, support needs, attack surface, and other burdens without comparative expansion of benefits. There is also need for comment from disciplines when it comes down to specific capabilities of software. For instance, understanding the difference between different statistics programs will not be possible for people who do not have nuanced experience with the software.

Usage and renewal evaluation

We do not currently have defined owners who have set specific goals for academic software nor do we have defined and regular reviews of data to evaluate progress towards those goals. This is a significant issue with major enterprise software like Canvas (the learning management system) and the Adobe Creative Cloud contract. It is also important to have similar evaluations  in place for discipline-specific software. These processes will vary greatly depending on the goals and availability of supporting data.

Without these goals and review processes, choices will continue to be made without sufficient participation by stakeholders and without sufficient attention paid to the actual contribution of the software to achieving pedagogical and research goals.