Which Database Do You Cut? Cost per Session Versus Obligation
Coastal Regional Middle/High School (890 students, grades 6–12) has $15,000 for digital resources next year. Three subscriptions are up for renewal: (A) a general periodical database at $6,200 with 4,100 sessions last year; (B) a science and math database at $5,900 with 780 sessions, almost all generated by four teachers' required research assignments; and (C) an e-book and translation platform at $4,300 with 2,300 sessions, of which 62% came from multilingual learners and students with read-aloud accommodations. Together they cost $16,400 - $1,400 more than the budget. A vendor also pitches a $3,100 test-prep product "fully aligned to state standards." The principal, who has the usage spreadsheet you gave her, says: "Cut the science one - look at those numbers - and use the savings for test prep." Your state library consortium offers a free general periodical package, but it lacks the backfile teachers cite.
1. What else do you need to know before agreeing, and how do you evaluate value beyond session counts?
Read the worked answer
Session counts measure captured audience, not value. Before agreeing I would need to know, for each resource: which courses require it, whether an equivalent is free through the state consortium, whether it is the accessibility route for students with read-aloud or translation accommodations, and how much usage comes from a handful of teachers versus across the school.
Applied to the numbers: A is roughly $1.51 per session and has an obvious substitute - the consortium package - with the backfile as the gap, which I would quantify by checking which cited articles actually came from it. B is about $7.56 per session, the worst ratio on the page and the one the principal flagged, but its 780 sessions come from required assignments in four courses. That could mean the resource is essential to required coursework, or that only four teachers know it exists; the spreadsheet cannot tell me which, and it cannot tell the principal either. C's $1.87 per session understates its value, because 62% of use comes from multilingual learners and students with read-aloud accommodations - an equity and legal obligation, not a popularity contest.
The question I would put back to the principal is not "which is lowest?" but "which of these can a student not do without, and is there a free replacement for the rest?" Cost per session is one column in the matrix; required curriculum, accessibility obligation, and substitutability are the others, and none of them appear in what she is holding.
2. What do you recommend, and how do you present it given that the principal has already given you a direction?
Read the worked answer
Recommendation: drop A, since the consortium package covers current-year content and I can route backfile requests through interlibrary loan at near-zero cost; that frees $6,200 and closes the $1,400 gap with room to spare. Then run B for one more year under a usage condition with a promotion plan, and do not add the $3,100 test-prep product in this cycle - "aligned to state standards" is a marketing claim, and I would want a 30-day trial with a defined group before any purchase.
But I would not simply overrule the principal with a spreadsheet of my own. She asked for a cut based on the only data she had; the professional move is to hand her a better matrix and make the trade-off explicit. In one memo: "Cutting B saves $5,900 and removes a database that four required courses assign from. Cutting A saves $6,200 and costs us a backfile we can borrow around. I recommend A, and I will take responsibility for the interlibrary loan volume."
If she still wants test prep, I would propose funding it from a negotiated reduction on B, not from C, and I would put both my recommendation and her decision in the same memo so the reasoning is documented at next year's renewal.
3. Design a short test to determine whether low usage of the science database reflects a poor product or poor promotion.
Read the worked answer
Six weeks of deliberate promotion separates a bad product from a well-kept secret. Design: pick two grade-level teams teaching comparable research units. Give one a 10-minute hands-on session tied to their actual assignment; give the other the usual email announcement. Compare authenticated sessions before and after, and set the success criterion in advance - for example, sessions from the targeted team triple and at least 60% of students in the targeted classes authenticate at least once - so the result cannot be rationalized afterward.
During the test I would collect qualitative evidence too: what teachers say when the database fails to answer their assignment, whether the interface works on the Chromebooks, whether the reading level fits the students assigned to it. And I would count authentications rather than page views, because a gatekeeper page can inflate a session count without a single real search.
If promotion works, the resource stays and the finding is about my outreach, not the vendor. If six weeks of genuine integration produces nothing, I have a defensible cut and a documented story for the principal, for the vendor at renewal, and for the teachers who will ask why it disappeared. Either way the decision stops resting on a raw session count.