Teacher Evaluation and Observation Software: What Happens When the Rubric Fits but the Bargained Timeline Does Not

The grievance that starts in April and ends in a settlement

It is the last week of April. Your office is assembling summative packets for 1,900 certificated staff. A middle school principal carries thirteen teachers on his caseload and has logged eleven of the required unannounced visits. Two are missing, and one of those two teachers is on a non renewal recommendation. The association representative asks one question: show me the dates. What you can produce is a record entered on April 19 for a walkthrough the principal says happened in November, a calendar invite that was moved twice, and an email thread. The rating may be entirely correct on the merits. It is not defensible on the record, so the district settles, the teacher stays, and every principal in the building learns that the process is theatre.

The money in this category is not the software line item. It is arbitration exposure, settlement cost, and the two to three weeks of HR and principal time that vanish every spring reconstructing a paper trail that should have been generated as a byproduct of doing the work. Across district projects we have delivered, the recurring pattern is the same: the rubric is fine, the forms are fine, and the process falls apart on dates, notice, and evidence.

Your collective bargaining agreement is the specification, and it is renegotiated

Everything that makes teacher evaluation hard is contractual, not pedagogical. How many announced and unannounced observations by tenure status. Whether a walkthrough under a stated number of minutes counts toward the required total. How many school days, not calendar days, may pass between the visit and the written feedback. How much notice a pre conference requires. Who is permitted to observe, and whether a second evaluator is required when a rating drops. When the improvement plan must be issued and how long it runs before a decision point. What the teacher's response window is and where the appeal goes.

Then multiply that by bargaining unit. Classroom teachers have one rubric. Counselors, psychologists, speech pathologists, and nurses usually have another with different domains and often different observation counts. Administrators have a third. Some states set the weight of student growth in the composite, and that weight has changed more than once in the last decade. Every one of those parameters is negotiable, and a settlement signed in July has to be running in September.

What Frontline, Standard for Success, Whetstone and PowerSchool Perform actually leave on the table

These are real products and none of them are bad. Frontline Professional Growth is strong on forms, workflow routing, and professional development tracking, and it sits inside a suite districts already own. Standard for Success is genuinely good at fast mobile observation capture with evidence tagged to indicators. Whetstone Education is the best of the group at coaching cadence and instructional leadership, which is the thing that actually improves teaching. PowerSchool Perform benefits from living next to the student information system, so rostering and staff records arrive without a nightly file.

The gap districts describe to us is consistent. First, none of them treat the contractual timeline as an enforceable constraint with school day arithmetic, per unit notice rules, and blocking behaviour. They record what you did; they do not stop you from doing it late or tell the principal on March 3 that he has eleven school days left and two visits outstanding. Second, composite formulas outside the common patterns require vendor configuration cycles, which means a contract settled in July may not be reflected in the tool until the year is underway. Third, evidence is stored, but it is not treated as a legal artefact: amendment history, lock points after teacher acknowledgement, and exact recomputation of a three year old rating under that year's formula are not what these products were designed to guarantee. If you have never had a rating grieved, none of that matters. If you have, all of it does.

Evidence is the artefact. The score is a byproduct

The defensible unit is not the rating, it is the observed statement with a timestamp, an author, an indicator tag, and a teacher response. A custom build makes capture the fastest path: a principal standing at the back of a room types low inference notes on a phone, drags each note onto rubric indicator 3b or 2c, and the system stamps the visit start and end from the device clock rather than from whenever the form was finally submitted. Artefacts such as lesson plans, student work samples, and family communication logs attach to the same record from the teacher's side.

Behind that, the store is append only. Corrections are recorded as amendments with the original preserved and the author named, and nothing is silently overwritten after the teacher acknowledges. That single design decision is what turns a hearing from a two day reconstruction into a printed timeline. It also changes principal behaviour, because everyone knows the record is the record.

The composite nobody can compute in a spreadsheet twice the same way

The summative rating is where districts quietly lose confidence in their own numbers. Observation domains carry weights. Student growth enters through a state produced measure for tested subjects and through locally written student learning objectives for everyone else, which is most of the staff. Some teachers share attribution across sections or co taught classes. There are minimum data thresholds below which a growth measure cannot be used, rounding rules, and cut scores that decide which band a teacher lands in. Change any one of those and half your ratings move a band.

What a custom build does is version the formula by school year and by bargaining unit, store the inputs rather than only the output, and expose a recompute function. When a teacher appeals a 2019 rating in 2021, you rerun the 2019 formula on the 2019 inputs and get the identical number, with a printable derivation showing every term. That is not an exotic requirement. It is simply not what a configuration screen in a packaged product is built to promise.

Improvement plans and the licence hours nobody is tracking

A rating below the threshold triggers a plan with a contractual duration, mandatory support activities, checkpoint dates, and often a second evaluator. Districts run these in Word. The plan then has to connect forward to professional development: the goal, the learning the teacher actually completed, the hours it carried, and the licence renewal cycle that consumes those hours. Most districts hold that in three unrelated places, which is why teachers arrive at renewal with a shoebox and HR spends June signing things it cannot verify.

Connecting the three is unglamorous work with an obvious payoff. The teacher sees one page: my goals, my observations, my plan, my hours, my renewal date. HR sees a queue instead of a shoebox.

What a custom build has to include

• Process templates scoped by bargaining unit and tenure status, with observation counts, notice rules, and windows expressed in school days against your actual district calendar
• Offline capable mobile observation capture with device timestamps and indicator tagging
• An evidence library where teachers and evaluators both contribute, with amendment history and lock on acknowledgement
• Composite scoring versioned by year and unit, storing inputs and supporting exact recomputation
• Student learning objective workflow with approval, mid year check, and attainment scoring
• Principal caseload dashboards that surface days remaining, not just work completed
• Improvement plan generation with checkpoints and second evaluator assignment
• Professional development catalogue with an hours ledger tied to licence renewal cycles
• State reporting export in whatever file layout your education agency demands
• Human resources and student information system synchronisation for staff, assignments, and rosters

What this costs and how long it takes

Across the 2,000-plus projects Digital Heroes has delivered, this is the honest shape. A first release covering observation capture, evidence, the timeline engine, and principal caseload dashboards runs $70,000 to $140,000 and ships in 12 to 18 weeks. That is a system your principals use in the first observation window, not a pilot. The full platform adding composites, student learning objectives, improvement plans, professional development hours, and state reporting runs $180,000 to $450,000 phased across 8 to 14 months.

What drives the number up: the count of distinct bargaining units and rubrics, because each is a separate process model rather than a copy. The state reporting file, which is usually a fixed width layout with validation rules that take real weeks. Growth measure mathematics, if your state uses a value added or growth percentile model you must ingest and align. Offline mobile, if principals observe in buildings with poor wireless coverage, which is more of them than anyone admits. Single sign on and human resources integration, particularly if the district runs an older on premise system. What keeps the number down: starting with the classroom teacher unit only, and one school year, then adding units in year two.

Build versus buy, stated plainly

Buy if you have one bargaining unit, one rubric, fewer than about 600 certificated staff, and a contract that has been stable for several cycles. Standard for Success or Frontline Professional Growth will serve you and a custom build would be an expensive way to own a maintenance burden.

Build when two or more of these are true. You run three or more distinct rubrics or bargaining units. You have more than roughly 1,500 certificated staff, at which point spring reconstruction becomes a staffing problem. You have lost or settled a grievance on process rather than on substance. Your state mandates a composite your vendor cannot configure without a change request. Your professional development and licence renewal tracking is a separate broken system. Or you are a state agency, charter network, or regional service centre operating this on behalf of many districts, in which case multi tenancy alone rules out most packaged options.

How to choose a developer for this

Ask them to model a school day calendar in front of you. If they reach for calendar days, or cannot immediately name the problem of a snow day inside a five day feedback window, they have not built this. Ask how they version a rubric mid year when a settlement lands in October and applies retroactively. Ask what happens to a submitted observation when the evaluator wants to change a word after the teacher has acknowledged it; the right answer involves an amendment record, not an edit.

Ask specifically about personnel record confidentiality. Evaluation data is an employee record, not a student record, and the access rules differ: a principal sees their caseload, HR sees the district, and a superintendent may or may not see a draft depending on your contract. Any developer who talks only about FERPA has misread the problem.

Ask who owns the code, the repositories, and the cloud accounts, and get it in writing before kickoff. At Digital Heroes the client owns everything from the first commit, and you should walk away from anyone who hedges on that question.