Keep reading.
Drop your email to finish the piece. One-time — we'll remember you on this device.
© 2026 Swiftscore, Inc.
Kelly Espinosa runs Fanny Jackson Coppin Elementary in Philadelphia. Four evaluations a week, 30 minutes each, plus coaching plans. Here is what changed when she stopped writing them from scratch.
CTO of Swiftscore. Writes about what we're learning from districts on the ground.
Drop your email to finish the piece. One-time — we'll remember you on this device.
Simplifies evaluation cycles and supports equitable teacher growth across multiple campuses.
Kelly Espinosa, principal of Fanny Jackson Coppin Elementary School in Philadelphia, was spending 30 minutes per teacher evaluation writeup and another two hours preparing the coaching conversation that followed. With four evaluations required each week, the math worked out to most of a workday gone before she got to a single classroom. After moving her observation notes through Swiftscore, writeup time dropped to two to three minutes per teacher, and coaching prep fell from two hours to fifteen minutes. The hours came back into classrooms.
Fanny Jackson Coppin Elementary is a Philadelphia School District K-8 building serving roughly 520 students. Kelly leads the building alongside two instructional coaches, Melanie and Marc. The school focuses on developing high-achieving learners and improving teacher effectiveness across the staff. Kelly's stated job is head of instruction. Her actual day is constant interruption: operations, facilities, student needs, the office phone.
That gap, between what a principal's job is on paper and how it actually fills the week, is the operating reality this case study is about.
Kelly is required to complete four formal teacher evaluations per week. Before Swiftscore, each one took at least 30 minutes of writeup time. The coaching plan that needed to come out of each observation added roughly two more hours per teacher. Stacked across the staff, the writing alone consumed days.
The frustration was not subtle.
"I literally will do everything else to avoid typing them up. I can't stand it. It's the least favorite part of my job."
Kelly Espinosa, Principal, Fanny Jackson Coppin Elementary
Two things compounded the problem. First, the document had to land aligned to school priorities and to the framework the district uses, not just be a transcript of what Kelly saw. Second, the coaching plan had to be specific enough that a teacher could actually do something with it on Monday. Both demands rewarded careful writing. Neither demand was compatible with a calendar broken into 11-minute fragments.
Kelly's workflow now starts the same way it always has: she sits in a classroom and takes observation notes by hand and on her laptop. The change happens after.
She runs those notes through Swiftscore. The tool produces a draft evaluation aligned to school priorities, separated into glows (strengths) and grows (areas for improvement), and pairs it with a coaching plan broken into weekly steps. Kelly reads it, edits where she wants to, and sends it. The observation evidence stays linked to the writeup, so when she pulls up the same teacher three months later she is reading against a history, not against a blank page.
What Kelly emphasizes is not the speed. It is that the draft is close enough to her own thinking that she does not have to fight it.
"I don't know how you guys got it that good."
Kelly Espinosa
The before-and-after, with no rounding up:
| Activity | Before | After | Change |
|---|---|---|---|
| Evaluation writeup, per teacher | 30 minutes | 2 to 3 minutes | ~93% reduction |
| Coaching prep, per teacher | 2 hours | 15 minutes | ~87.5% reduction |
| Formal evaluations per week | 4 | 4 | unchanged |
| Time back in classrooms per week | baseline | ~3 to 5 hours | recovered |
The "time back in classrooms" line is the one Kelly cares about. The other rows are how it got there.
A case study is only useful if you can see the work. Here is the kind of growth Swiftscore generated from one of Kelly's writing-lesson observations:
"Instead of recall-based questions, incorporate higher-order thinking questions."
It also offered the teacher a concrete prompt to use the next day:
"How can you use evidence to build a strong point in your essay?"
Kelly described the coaching plan as "bite-sized" and "specific." That is the bar a coaching plan has to clear to actually change practice. A teacher cannot work on "improve questioning." A teacher can swap a recall question for an evidence question on Tuesday.
The temptation, when you cut a 30-minute task down to three minutes, is to celebrate the 93% number and stop there. The interesting question is what happens to the reclaimed hours. In Kelly's case, the answer is that they go back into the part of the job that matters: being in classrooms, observing more often, and meeting with coaches Melanie and Marc on what they are seeing across the building.
Four formal evaluations per week is a floor, not a ceiling. With writeup compressed, the constraint on how often Kelly walks into a classroom is no longer paperwork. It is her own calendar.
There is a secondary effect worth naming. When evaluations are easier to finish, they are also easier to finish on time. Teachers get feedback in the same week they were observed, not three weeks later when the moment is gone. That alone changes how feedback lands.
We are tracking three things in Kelly's building over the rest of the year:
The first two we will know by mid-year. The third is a longer arc. We will write about each of those when there is something honest to report.
For now, the line that sums up where Kelly is:
"I would be knocking down your door saying, 'Where's Swiftscore?' If the district won't pay for it, I'll pay for it myself."
Kelly Espinosa
Swiftscore aligns the generated draft to the framework the district uses. Kelly's school uses Philadelphia School District's evaluation framework. The tool also supports Danielson and Marshall rubrics, among others, and respects school-specific priorities a principal sets at the building level.
Yes. Swiftscore generates a first draft from the principal's observation notes. The principal reads, edits, and sends. The judgment stays with the principal. The typing does not.
For Kelly's building, first usable evaluation came out of the tool in the same week she started. There is no multi-month implementation.