User guide & methodology
eLearning Development Estimator
A practical guide to creating, testing, and communicating a defensible eLearning development estimate.
1. Quick start
- Define the scope. Enter the project name, finished learning time, modules, complexity profile, and delivery characteristics.
- Check the effort model. Select or edit the development-hour ratios and phase allocation. Add media, accessibility, localization, and other project work.
- Model the delivery team. Enter the blended hourly rate, weekly productive capacity, and review/approval wait time.
- Compare AI responsibly. If enabled, enter realistic reductions, direct tool costs, and added human review or remediation.
- Review the evidence. Resolve warnings, document assumptions, and use the readiness prompts before sharing.
- Open the report. Print it, save it as a PDF, or save the inputs as JSON for later revision.
Fastest orientation: choose Example in the estimator toolbar, inspect all five steps, then select Reset to start your own estimate.
2. Scope and project profiles
Finished learning time
Enter the total learner-facing runtime, not the development team's work hours. The estimator converts minutes to finished hours before applying the selected development ratio.
Modules and level
Modules help describe packaging and can affect add-on work. Choose the level that best reflects the expected instructional and production complexity—not the label that produces the preferred budget.
Lean, typical, and complex profiles
These are planning profiles—not confidence intervals and not statistical predictions. They represent different delivery conditions. Lean assumes favorable conditions and reuse; typical is a reasonable working case; complex/media-heavy reflects greater production or stakeholder burden.
| Level | Lean | Typical | Complex / media-heavy |
|---|---|---|---|
| Level 1 | 49:1 | 79:1 | 125:1 |
| Level 2 | 127:1 | 184:1 | 267:1 |
| Level 3 | 217:1 | 490:1 | 716:1 |
Important: the default ratios come from Chapman Alliance research published in 2010. They predate current AI-assisted workflows and are broad reference points. Replace them with comparable internal project data when you have it.
3. Effort, cost, and timeline assumptions
Phase allocation
The phase percentages distribute core development effort across analysis, design, development, quality assurance, and project management. If your entries do not total 100%, the estimator normalizes them for calculation and shows a warning.
Add-ons and localization
Use add-ons for work not already represented by the base profile: custom video, animation, graphics, accessibility remediation, LMS work, or similar items. Avoid double-counting. Advanced ratios may already imply substantial media production.
Localization is applied to core build effort for each additional language. Custom media is estimated separately because dubbing, re-editing, subtitles, and localized graphics can vary significantly.
Rate and team capacity
The blended hourly rate converts effort to estimated labor cost. Weekly productive capacity converts hours to active work weeks. Use true project capacity—not nominal paid hours—after meetings, administration, competing assignments, and planned leave.
Review and approval wait
Add expected calendar time for stakeholder review and approvals. This makes the timeline more realistic, but it still does not create a resource-loaded project schedule.
4. AI-assisted comparison
The AI scenario is a transparent comparison, not an automatic savings claim. It applies entered reductions to eligible design and development work, then adds direct tool/vendor costs and human review or remediation hours.
- Design and development reductions: enter only reductions supported by your workflow, content type, governance, and team capability.
- Voice and localization options: model these independently when AI materially changes that work.
- Tool/API/vendor cost: include licenses, usage fees, external services, and other direct scenario costs.
- Human review/remediation: include fact-checking, editing, accessibility checks, brand review, testing, privacy review, and rework.
Unspoken truth: AI does not guarantee a cheaper project. Weak inputs, high-risk content, extensive governance, or heavy remediation can erase the apparent labor savings. The estimator intentionally allows the AI scenario to cost more.
5. Results and reporting
Review the hours, labor cost, direct AI cost, calendar duration, phase allocation, and scenario differences together. A low total is not automatically a good estimate; it may simply mean important work is missing.
Before sharing the report
- Resolve or explain every warning and evidence gap.
- Confirm that the chosen profile matches the proposed learner experience.
- Check for double-counted or omitted media, accessibility, localization, LMS, and review work.
- Validate rate, capacity, and approval time with the people responsible for delivery.
- State what is excluded and identify the assumptions most likely to change.
Select Report in the estimator toolbar to open the presentation view. Use Print / Save PDF for a portable copy and Save inputs to preserve the editable assumptions.
6. Calculation methodology
Finished hours = finished learning minutes ÷ 60
Core development hours = finished hours × selected development ratio
Total effort = core hours + localization + custom media/add-ons + AI review/remediation
Estimated cost = labor hours × blended hourly rate + direct AI/tool/vendor cost
Active work weeks = total labor hours ÷ weekly productive team capacity
Calendar estimate = active work weeks + review/approval wait weeks
The estimator evaluates lean, typical, and complex/media-heavy ratios separately. The phase chart uses normalized phase percentages and includes separately estimated add-ons so its total reconciles to the selected scenario's total hours.
7. Saving, loading, and privacy
- Automatic browser save: inputs are stored in this browser's local storage so they can persist between visits on the same device and site address.
- Save: downloads a JSON file containing the editable assumptions.
- Load: imports a compatible estimator JSON file after validating its format.
- Reset: clears the estimator's saved browser state and returns to a blank model.
The estimator runs in the browser and does not intentionally transmit project inputs to Cahill Consultants. Your website host, browser, extensions, analytics, or organizational controls may still process normal page-use data; review your own hosting configuration before making broader privacy claims.
8. Limitations and sources
What the estimator does not model
- Task dependencies, resource contention, individual assignments, or fixed launch-date scheduling
- Maintenance, future revisions, hosting, learner support, or post-launch measurement unless you add them explicitly
- Statistical confidence, market pricing, vendor quotes, or guaranteed AI productivity
- Every governance, legal, privacy, accessibility, or regulated-content requirement
Reference sources
- Chapman Alliance (2010), How Long Does It Take to Create Learning?, source of the default 49/79/125, 127/184/267, and 217/490/716 ratios. A reproduction is available in this Learning Guild presentation.
- ATD: How Long to Develop One Hour of Training? Updated for 2017, which emphasizes that format, tools, templates, and project context materially affect development time.
Document the date, evidence, and owner for every local calibration. Over time, compare estimates with actual effort and revise the assumptions—the tool becomes top-tier only when the organization treats estimation as a learning system rather than a one-time calculation.