Estimation sounds scientific. We wrap it in spreadsheets, formulas, and confidence intervals.
Underneath all of that, an estimate is still a guess. The moment you swap the word
estimate with guess,
our language gets uncomfortable:
-
“Can you give me an accurate guess?”
-
“We need a precise guess by Friday.”
-
“Please refine your guess and bring it to steering.”
We reach for the word estimate because guess feels unprofessional. The truth is simpler:
an estimate is an informed guess based on what we know at a specific point in time.
The only number we know for certain is the one we see when the project is over.
Why estimates improve as we go
Project work is “progressively elaborated.” The more we do, the more we see. The more we
see, the better our judgment. The better our judgment, the more informed our estimates.
Imagine the lifecycle as a funnel. At the top, your knowledge is thin and your estimate
range is wide. As you move towards completion, experience narrows that funnel. You’ve seen
real issues, real fixes, and real constraints. Accessibility work follows the same pattern
as security, performance, or architecture.
The Polaris lesson: estimating the unknown
In 1957, the United States Navy faced a familiar problem. They were developing the Polaris
missile—the first solid-fueled, nuclear-armed, submarine-launched ballistic missile. They
had never built anything like it and still needed a timeline and a budget.
Waiting until the end to “find out” wasn’t an option. So they created the
Project Evaluation and Review Technique (PERT) to
estimate work with limited history and high uncertainty. Today, PERT still solves the same
problem we face when we look at accessibility findings for the first time.
PERT in one glance
For each activity, you imagine three outcomes:
- Best case (bc): shortest realistic time.
- Worst case (wc): longest realistic time.
- Most likely (ml): expected time under normal conditions.
PERT combines them like this:
Estimate = (bc + 4 × ml + wc) / 6
It doesn’t make the estimate “correct.” It makes it structured, explainable, and repeatable.
Applying PERT to accessibility work
A good accessibility assessment tells you what was found, how often it appears, where it
lives in your code, and how to fix it. That means you’re not estimating “accessibility
magic.” You’re estimating familiar work: code changes, design decisions, and QA checks.
Suppose your report highlights three issue types:
-
Name, Role, Value
-
Information and Relationships
-
Color Contrast
Each behaves differently when you estimate effort. Name, Role, Value issues often look like
straightforward code fixes. Information and Relationships might need design and UX input.
Color Contrast can show up as a long list of findings but come from a small number of
underlying tokens or variables.
A concrete example: Name, Role, Value
Take a Name, Role, Value issue from your audit where you have a code snippet and a
recommended fix. That means the work is already halfway known. Your developer and QA
analyst sit down to estimate one representative case:
- Worst case: “If everything goes wrong, maybe 40 hours.”
- Best case: “When it’s simple, we’ve done this in 2 hours.”
- Most likely: “Usually this type of fix takes about 6 hours.”
bc = 2, ml = 6, wc = 40
Estimate = (2 + 4×6 + 40) / 6 = 66 / 6 = 11 hours
You get an 11-hour estimate per Name, Role, Value issue. Multiply by the number of issues
of that type and you have a defensible starting point. As the team fixes the first few,
you refine the estimate based on experience rather than theory.
How Agile teams use Planning Poker for accessibility
Many Agile teams use Planning Poker or Story Points instead of hour-based estimates.
It looks different from PERT on the surface, but the purpose is the same:
take something unknown and make it comparable.
Planning Poker was never about precision. It was designed to surface uncertainty, expose
risk, and force conversation. That maps cleanly to accessibility work—especially when a
team believes they’ve “never done this before.”
With a Name, Role, Value issue from your audit, the unknown becomes known quickly:
you have the code snippet, the recommended fix, and a pattern that looks like other
markup or component refactors the team has already done.
Using PERT thinking inside Planning Poker
- Pick one representative accessibility issue with a snippet and a fix.
- Ask, “What past story does this feel like?” and anchor it to known work.
- Talk through best case, worst case, and most likely outcomes—even informally.
- Choose a Story Point that reflects that spread, not a single perfect number.
- Use that story as the baseline for similar issues in future sprints.
You don’t need to show the PERT formula in an Agile ceremony. You’re borrowing the same
mental model: acknowledge uncertainty, compare it to known work, and align as a team.
For predictable, pattern-based issues—like many Name, Role, Value changes or component-level
fixes—this approach stabilizes quickly. By the second or third sprint, accessibility stops
feeling like a foreign category of work and becomes another familiar stream in the backlog.
For more complex issues—such as Information and Relationships, Color Contrast that touches
your design system, or component restructuring—Planning Poker becomes an alignment tool.
Design, development, and QA have to be in the conversation so the Story Points reflect
scope and impact, not just effort.
PERT gives you a formula. Planning Poker gives you a conversation. Both exist to make
uncertain work discussable, comparable, and plannable. Accessibility estimation belongs
there as a first-class citizen.
From fixing issues to preventing them
Fixing accessibility issues is necessary, but it is not a strategy. Many organizations
treat accessibility as a one-time remediation event: run an audit, fix the issues, move on.
The real leverage comes when you shift from remediation to prevention:
-
Analyzing the root causes behind recurring issue types.
-
Updating design systems, patterns, and training so teams build it right the first time.
-
Embedding accessibility checks into everyday delivery, not just special audits.
The later you find a problem in the lifecycle, the more expensive it is to fix. By the time
an accessibility issue surfaces in production, you are paying in rework, backlog churn, and
customer trust. Prevention is not just good practice; it is good economics.
How this insight supports different learners
R
Readers
Get a clear mental model for turning accessibility estimates from guesswork into informed,
explainable numbers.
L
Listeners
Hear how Polaris, PERT, and Agile ceremonies all tackle the same problem: making the
unknown plannable.
D
Doers
Walk away with a repeatable way to estimate Name, Role, Value and other issues—whether
you use hours or Story Points.
O
Observers
See how shifting from remediation to prevention changes the conversation with leadership
about accessibility and risk.