A polished EdTech demonstration can make three different products look equally persuasive. Each promises personalization, helpful dashboards, easy setup, and stronger learning. The first has an impressive AI assistant. The second offers unusually smooth integration. The third costs less per student. Which should a school choose?
Compare EdTech vendors by the learning problem they can solve, the quality and relevance of their evidence, the real work required to deploy them, and their total value over several years—not by their feature lists alone. A useful comparison puts vendors against the same educational outcomes, requires the same demonstrations and documentation, calculates the complete cost of ownership, and applies a decision rule agreed upon before sales presentations.
This article is a decision framework for US K–12 superintendents, principals, curriculum directors, technology leaders, purchasing teams, and instructional coaches. It is about choosing between competing suppliers, not simply deciding whether one product deserves a pilot. For a complementary checklist of questions to ask before a pilot, see .
Start with the learning need, not the vendor shortlist
Imagine a district where seventh-grade students can perform routine calculations but struggle to explain mathematical reasoning. Teachers want timely feedback on misconceptions, and the district has a limited budget for supplemental practice. A vendor begins its presentation with an animated dashboard, dozens of AI features, and hundreds of ready-made exercises.
The procurement team should interrupt its own enthusiasm with a simpler question: What would improve for students if we bought this?
A meaningful needs statement names the learner group, the specific instructional difficulty, the existing approach, and the observable improvement sought. For example:
Our Grade 7 mathematics teams need a manageable way to identify and respond to misconceptions in proportional reasoning while improving students’ ability to explain unfamiliar problems independently.
That wording immediately changes the comparison. A platform that assigns more questions is not automatically better. A product that identifies misconceptions but overwhelms teachers with unusable data may not help. And a low-cost program might be the strongest choice if it supports thoughtful feedback without adding a second workflow.
The Digital Promise EdTech Procurement Framework starts with analysis and inventory before discovery, implementation, evaluation, and scale. Its underlying lesson is important: purchasing should respond to a documented instructional need, not manufacture one around a promising product.
Before inviting vendors, inventory current tools. Determine which products already perform similar functions, how often schools actually use them, whether contracts overlap, and whether a change in classroom practice could solve the problem without buying another platform. This is not hostility to innovation; it is respect for scarce educational resources.
Define the success criteria in advance
Write down two or three outcomes that would justify the investment. The primary outcome should concern learning: perhaps improved performance on independently scored explanation tasks or better retention on a delayed assessment. Secondary outcomes might include teacher workload, accessibility, reliable rostering, and student confidence.
Specify what evidence would count as progress and which trade-offs the district will not accept. For instance, a district may refuse a vendor that cannot meet mandatory privacy, accessibility, or security conditions regardless of its instructional score. That boundary must exist before anyone sees a compelling demonstration.
Compare all vendors with the same seven dimensions
A common error is to let each sales team define the comparison. Vendor A emphasizes personalization, Vendor B emphasizes analytics, and Vendor C emphasizes affordability. The school ends up comparing three marketing stories instead of three responses to one problem.
Use a single rubric that every supplier must answer. The following illustrative weights total 100 points; schools should revise them with instructional, technical, financial, and community stakeholders.
| Decision dimension | Example weight | What credible evidence looks like |
|---|---|---|
| Learning value and evidence | 30% | Relevant research, tasks, independent student outcomes |
| Classroom usability and teacher capacity | 15% | Realistic teaching workflow and support requirements |
| Technical integration and interoperability | 15% | Tested roster, identity, LMS and reporting connections |
| Privacy, security and accessibility | 15% | Completed specialist review and acceptable safeguards |
| Three-year total cost of ownership | 15% | All-in financial model with clear assumptions |
| Vendor reliability and contract terms | 5% | Service commitments, exit rights and transparency |
| Student independence and agency | 5% | Tasks that build understanding beyond the platform |
Use a non-negotiable eligibility gate before scoring. If a product fails a required legal, safeguarding, data-security, accessibility, or procurement condition, it cannot win merely because a weighted average is high. The weights help compare acceptable candidates; they do not make compliance optional.
Scoring should happen in two rounds. In the first, teams examine documentation and ask for evidence. In the second, shortlisted vendors complete the same realistic tasks using school-created examples with synthetic or otherwise approved data. Keep individual scores and written reasons before a committee discussion, so one confident presenter does not set the tone for everyone.
Make learning evidence the largest dimension
A claim such as “students improved by 35%” tells very little without context. Improvement in what? From what baseline? Compared with which alternative? Over how many weeks? On the product’s own quiz or an independent measure? In which student population?
Ask each vendor for the original study, not a slide summarizing it. Review the age groups, prior attainment, learning setting, implementation conditions, attrition, outcomes, comparison group, and whether researchers were independent of the vendor. Distinguish published, methodologically credible evidence from internal case studies and customer testimonials. Evidence from elementary literacy classrooms should not automatically be treated as evidence for secondary mathematics.
The US What Works Clearinghouse reviews research on educational interventions and provides methods for judging study quality. It does not endorse individual products or guarantee that a positive result in one setting will transfer to yours. Use it to think critically about the strength and limits of research, and record when meaningful evidence is absent.
A school can also ask whether the program has a plausible mechanism. Does it help teachers diagnose misunderstandings? Does it provide well-sequenced practice? Does it encourage students to retrieve knowledge, explain reasoning, and learn from feedback? Or does it mainly increase the number of tasks they complete?
More usage is not synonymous with more learning. Where AI is involved, inspect whether learners can complete analogous work without generated hints. If the tool provides the entire solution, a high completion rate may reflect assisted performance rather than independent competence.
Test technical integration as a real classroom process
“Integrates with your LMS” is not a complete answer. An advertised connection can still require daily manual imports, produce duplicate records, lose assignment context, or create a separate support burden for teachers.
Ask each finalist to demonstrate a full, approved workflow from beginning to end: provisioning a class, signing in through the district’s identity system, assigning work, accessing results, correcting a roster change, and exporting the information needed if the contract ends. Use a temporary test environment and appropriately protected data.
Relevant 1EdTech interoperability standards include Learning Tools Interoperability (LTI) for connecting learning tools with platforms and OneRoster for exchanging K–12 class and roster information. Standards support interoperability but do not, by themselves, prove that a particular deployment works. Verify the supported versions, whether relevant certifications apply, the direction of data exchange, and what has actually been tested.
A teacher might be told that a product saves ten minutes per class. That promise has little value if getting a roster into the system requires an administrator every Monday or if feedback arrives after teachers have already planned the next lesson.
Ask the questions IT and curriculum teams ask differently
The IT team may want to know whether the product supports single sign-on, encryption, logging, incident reporting, stable APIs, data retention controls, and practical support. The curriculum team may ask whether the tool aligns with the intended content, provides actionable feedback, and fits the lesson sequence. Special education and accessibility staff will look for usable alternatives and assistive-technology compatibility. Teachers will ask how the tool changes the classroom minute by minute.
All those perspectives matter. A cross-functional committee should combine them without allowing the most technical or the most pedagogically enthusiastic voice to displace the others.
Build a test script in advance. Give every vendor the same hypothetical Grade 7 class, the same instructional objective, and the same questions. Ask for a live walkthrough of an ordinary week—not a rehearsed flagship lesson. Record where extra clicks, workarounds, or manual cleanup appear.
Put privacy, accessibility and security through separate gates
Data protection is not a bonus feature. Schools need to understand what student information an application collects, why it collects it, who can access it, how long it is kept, whether it is used for model training or advertising, how incidents are handled, and how data will be returned or deleted.
The US Department of Education’s Student Privacy Policy Office guidance advises teachers to check with school or district administrators before using services that may involve protected student information. FERPA and other applicable requirements need to be considered by qualified district personnel; this article does not offer legal approval for a particular tool.
A supplier’s generic “FERPA compliant” badge is not a substitute for examining the district’s contractual and operational responsibilities. The 1EdTech Data Privacy resources offer an additional vetting framework, but a seal does not replace local judgment.
Accessibility deserves an equally practical review. Do not merely request a PDF and file it away. Ask users with relevant accessibility expertise to test keyboard navigation, screen-reader support, contrast, captions, alternative formats, error messaging, and realistic tasks. The appropriate district team should decide the applicable requirements and whether the vendor’s documentation matches the deployed product.
If a supplier cannot make basic safeguards transparent, it is not yet ready for a normal weighted comparison. First resolve the gate; then score instructional fit.
Calculate three-year total cost, not just price per seat
The cheapest license may become the most expensive product to sustain. Schools pay through more than invoices: they also pay through educator training time, IT support, device constraints, integration work, contract administration, and disruptions caused by unreliable workflows.
CoSN’s Total Cost of Ownership resources distinguish hardware, software, infrastructure, support, training, and indirect staff labor. That is a more useful basis for comparing long-term value than the headline subscription fee.
For each vendor, request a written three-year cost model that includes at least:
Licenses, minimum seat commitments, renewal increases, and any required add-ons.
Setup, data migration, integration, and implementation fees.
Initial and recurring professional learning, including replacement-staff onboarding.
Expected internal IT, curriculum, helpdesk, and administrative hours.
Devices, connectivity, accessibility adjustments, and technical dependencies.
Support tiers, upgrades, escalation fees, and contract management.
Exit costs: data export, migration, staff retraining, and overlap with a replacement.
Separate actual quotes from estimates. Include low, expected, and high scenarios where use and staffing are uncertain. Do not assign a speculative dollar value to learning outcomes just to make a weak business case look precise.
A transparent example of why the subscription price can mislead
Consider two hypothetical vendors serving 1,000 students. The figures below are illustrative arithmetic, not real vendor prices or procurement findings.
| Three-year cost component | Vendor A | Vendor B |
|---|---|---|
| Licenses | \$45,000 | \$60,000 |
| Implementation and integration | \$12,000 | \$3,000 |
| Training and support labor | \$24,000 | \$9,000 |
| Additional infrastructure and transition | \$9,000 | \$3,000 |
| Estimated three-year total | \$90,000 | \$75,000 |
| Three-year cost per targeted student | \$90 | \$75 |
Vendor A looks cheaper in the first row but costs more in the complete model. Neither total tells us which product teaches better. Cost is a decision dimension, not a proxy for learning.
Ask what happens if usage is only half of what was projected. Can the district reduce seats? Are licenses transferable between schools? What are the renewal terms? Will the vendor charge to retrieve data in a usable format? A realistic total-cost comparison should make expensive surprises visible before the agreement is signed.
Run an identical demonstration challenge for each finalist
Instead of asking “Can your platform personalize instruction?”, ask each vendor to solve exactly the same constrained problem.
Provide a synthetic set of student responses involving one common misconception and one more complex reasoning error. Ask the vendor to demonstrate how a teacher identifies those patterns, chooses an appropriate instructional response, and checks whether students can explain the concept independently one week later.
Observe how much of the workflow can actually be performed. Does the teacher need to inspect fifty dashboard rows manually? Can students articulate what they need to improve? Does the system merely recommend more content, or does it reveal something useful about misconceptions? Can the teacher override an unsuitable recommendation?
Then use a second scenario: a new student joins midterm, a teacher is absent, and two students need accessible alternatives. The platform that performs beautifully in a curated demonstration may struggle when school routines become ordinary and messy.
Each team should record observations against the rubric rather than general reactions such as “intuitive,” “innovative,” or “impressive.” Those adjectives may signal genuine strengths, but they do not describe the evidence.
Give teachers and students a structured voice
Teachers know which workflows are realistic. Students can help identify confusing instructions and whether a platform encourages independence or repeated hint-seeking. Invite feedback without treating popularity as a substitute for learning evidence.
A quick student or teacher feedback survey might ask: What helped you make sense of a difficult concept? When did the tool become an obstacle? What could you complete without assistance afterward? Which extra tasks did the tool create for you?
Use these responses as contextual information. Do not disguise a convenience sample as a formal scientific study, and protect students’ privacy and consent according to local requirements.
Use a weighted scorecard—and disclose its limits
A scorecard can make decisions more transparent, but a precise-looking number is not objective merely because it has decimals. Weights are value judgments. Ratings may be uncertain. Some strengths are not commensurable. Treat the tool as a way to organize deliberation, not as an automatic purchasing authority.
One simple method is to rate each eligible vendor from 0 to 5 on each dimension, write a short rationale, and calculate:
Weighted score = sum of (dimension rating ÷ 5 × dimension weight).
For illustration, imagine three eligible suppliers with these hypothetical results:
| Dimension (weight) | Vendor A | Vendor B | Vendor C |
|---|---|---|---|
| Learning evidence (30) | 4 | 3 | 5 |
| Teacher fit (15) | 3 | 5 | 3 |
| Integration (15) | 5 | 4 | 3 |
| Privacy/accessibility (15) | 4 | 4 | 4 |
| Three-year cost (15) | 2 | 5 | 3 |
| Contract reliability (5) | 3 | 3 | 4 |
| Learner independence (5) | 4 | 4 | 5 |
| Weighted total /100 | 73 | 79 | 78 |
These example scores are a teaching model, not test results. Vendor B leads narrowly. The committee should still ask whether its weaker evidence rating is acceptable, whether the apparent cost advantage survives contract review, and whether changes in the weighting would reverse the ranking.
Run a sensitivity test. What happens if learning evidence receives 40% rather than 30%, or if staffing costs are higher? A decision that flips under small reasonable assumptions deserves closer investigation. The right response may be a limited pilot or a request for better documentation rather than an immediate award.
Do not average away a decisive weakness. If a supplier fails an essential security review or cannot serve a required learner population, a high aggregate number does not cure that deficiency.
Negotiate the relationship, not only the subscription
Procurement is often treated as a one-time selection event. In reality, the contract defines the district’s relationship with a vendor for years.
Clarify service-level expectations, accessible support routes, escalation during incidents, changes to product features, pricing adjustments, data ownership, use of de-identified information, subcontractors, and the process for ending the agreement. For AI-enabled tools, ask how significant model or feature updates will be communicated and re-evaluated. A product tested in February may not behave identically after a material update in October.
Plan for teacher turnover and district restructuring. If the vendor’s account manager departs, can the school still understand how the program works? If one principal champions the product and later leaves, does the value proposition survive? Sustainable learning tools should be understandable and governable by the institution using them.
An exit plan is not evidence that a relationship is expected to fail. It is how a district protects continuity for students and teachers if circumstances change.
A 30-day vendor-comparison process for school leaders
The following schedule is a suggested administrative workflow, not a validated procurement timeline or substitute for local bidding rules.
Days 1–5: define and prioritize
Assemble a small team from curriculum, teaching, IT, data privacy, accessibility, and procurement. State the precise learner need and review existing licenses. Agree on required safeguards, seven comparison dimensions, weights, budget limits, and the evidence every supplier must provide.
Days 6–12: request comparable submissions
Send the same written brief to each supplier. Ask for research reports, a typical classroom workflow, system architecture and standards support, accessibility documentation, privacy terms, implementation staffing, references, and a three-year all-in quote. Require disclosure of assumptions and exclusions.
Days 13–20: test realistic tasks
Conduct identical demonstrations. Have teachers walk through an ordinary lesson and IT staff test the approved integration environment. Record evidence, friction, and gaps. Invite structured feedback rather than unrestricted product voting.
Days 21–25: evaluate and challenge the results
Score each eligible vendor independently, then discuss differences. Run total-cost and sensitivity scenarios. Identify claims needing further verification. Review whether a smaller pilot, a non-digital solution, or an existing district tool may be preferable.
Days 26–30: decide and document
Record the rationale, uncertainties, approved conditions, responsible owners, and evidence to review after implementation. Choose to award, negotiate further, pilot a finalist, or purchase nothing. If procurement law or district procedures require a different sequence, follow those requirements.
The purpose is not to rush an award in 30 days. It is to make every stage observable and accountable.
The Learning Coach perspective: buy tools that make learners less dependent
Many vendor comparisons stop at return on investment, performance gains, or operational efficiency. Those questions matter, but education has an additional responsibility: to develop the learner’s capacity to act without constant guidance.
A strong learning tool should give educators better visibility into where understanding breaks down, enable meaningful feedback, and help students acquire strategies they can use on their own. It should not merely create a busier screen or produce completed assignments faster.
That principle echoes the argument in Stop Teaching. Start Coaching.: the ultimate measure of education is not how much information adults or systems deliver, but how capable students become. In a technology decision, ask what happens when the software is switched off. Can students still explain, recall, plan, solve, and evaluate their work?
This also changes the role of the school leader. A pedagogical leader does more than approve purchasing. They help teachers investigate the learning problem, interpret evidence, adapt classroom practice, and evaluate whether the change benefits students. The Learning Coach approach calls for observation, reflective questions, practical feedback, and progressive learner autonomy—not a promise that software alone will transform education.
Educators and school leaders who want to develop that professional approach can explore Learning Coach training and its specializations. The pathway and availability of individual specializations should be confirmed on the current application page.
Frequently asked questions
What matters most when comparing EdTech vendors?
Start with a clearly specified learning problem and relevant evidence that the proposed tool can address it. Then evaluate teacher fit, integration, privacy and accessibility, total cost, contractual flexibility, and student independence. No single feature should dominate without a compelling educational reason.
Should a school always select the lowest-cost vendor?
No. It should compare the full cost over a meaningful period and judge that cost alongside expected educational value and risk. A lower-cost product may be attractive if it meets the same learning need with less implementation effort. But no amount of apparent savings makes an unsafe or educationally unsuitable product a good purchase.
How many EdTech vendors should be shortlisted?
There is no universal number. Select a manageable group that satisfies preliminary eligibility rules and allows meaningful side-by-side evaluation under the district’s procurement procedures. More finalists are not necessarily better if staff cannot review the evidence properly.
Can a vendor's own research be trusted?
It can be informative, but methods, samples, outcomes, comparison conditions, and financial interests need scrutiny. Independent replication is valuable when available. Absence of rigorous research should be recorded as uncertainty, not automatically interpreted as proof of failure.
What if none of the vendors meets the learning need?
Purchasing nothing can be a responsible decision. Reconsider the original problem, adapt existing resources, strengthen teacher practice, or postpone the purchase until better evidence or better-fitting solutions become available.
A better final question for the purchasing committee
A selection committee can spend hours debating which platform is most innovative. A more useful final discussion is: Which acceptable supplier gives our teachers the most realistic path to stronger, more independent student learning at a cost and level of risk we can sustain?
If the answer is supported by consistent criteria, credible evidence, tested workflows, a complete financial model, and a written implementation and exit plan, the district has compared vendors rather than simply compared demonstrations.
Features describe what a product can do. Educational value depends on what learners and teachers can reliably accomplish because of it.
Resources
Digital Promise — EdTech Procurement Framework — six-stage framework for selection, piloting and sustainable adoption.
Institute of Education Sciences — What Works Clearinghouse — evidence reviews and research-quality resources.
1EdTech — Interoperability Standards — specifications including LTI and OneRoster.
1EdTech — Data Privacy — supplier vetting resources.
CoSN — Total Cost of Ownership — lifecycle technology cost methodology.
U.S. Department of Education — Student Privacy Policy Office — guidance on educational apps and FERPA considerations.
