Elimination by Aspects

TL;DR

Elimination by Aspects: When choosing among many options, rank your criteria by importance. Apply the most important criterion first — eliminate everything that doesn't meet your threshold. Apply the second criterion to what remains. Continue until one option remains. Simple, fast, and works well when one or two criteria dominate.

What Is Elimination by Aspects?

Choosing between many options with multiple attributes is cognitively overwhelming. A car buyer comparing 20 models on 15 attributes — price, fuel economy, reliability, safety rating, cargo space, brand, warranty — faces 300 data points to process simultaneously. Humans can't actually do this. What they do instead is use heuristics — shortcuts that reduce the cognitive load.

Amos Tversky formalized one of the most common of these shortcuts in 1972: Elimination by Aspects (EBA). The process: rank your decision criteria by importance. Apply the most important criterion first, eliminating all options that don't meet a threshold. Then apply the second criterion to the surviving options. Continue until one option remains.

EBA is not optimal in the technical sense — it can eliminate an option that is excellent on the second and third criteria but slightly subpar on the first, and it can miss that the "almost-good-enough" first-criterion option would have been the best overall choice. This is the key criticism of the model.

But Tversky's contribution was empirical: this is how people actually make decisions when faced with complex choices. Understanding EBA explains behaviors that look irrational from a classical utility-maximization perspective — and it offers a simple, disciplined framework for improving on the default version of this heuristic.

The critical improvement over the default: be deliberate about the order of your criteria and the thresholds you apply. Most people apply EBA unconsciously and inconsistently. Making it explicit turns a cognitive shortcut into a disciplined tool.

How It Works

Elimination by Aspects — Process

Step 1: List all viable options
  Don't filter yet — gather the full candidate set.

Step 2: Identify all relevant criteria
  What attributes matter in this decision?
  Write them all down without ranking yet.

Step 3: Rank criteria by importance (most important first)
  This is the critical step. Ask: if I could only consider one criterion,
  what would it be? Then: if I could add a second, what?
  Be honest — don't list "brand prestige" first if price really matters most.

Step 4: Set a threshold for the first (most important) criterion
  What is the minimum acceptable level?
  This is a pass/fail threshold, not a ranking.

Step 5: Eliminate all options that fail the first threshold
  Any option below the threshold is removed from consideration.

Step 6: Repeat with remaining options on the second criterion
  Apply the second criterion threshold to surviving options.
  Eliminate all that fail.

Step 7: Continue until one option remains, or until the set is
  small enough to evaluate holistically

Recovery step if all options are eliminated:
  Relax the threshold on the least important criterion you've applied,
  or revisit whether the threshold was appropriate.

Three Real-World Examples

Hiring: Filtering a Candidate Pool

A hiring manager receives 200 applications for a senior engineer role. She identifies her criteria in order: (1) at least 5 years of backend experience — minimum threshold; (2) experience with distributed systems — minimum threshold; (3) evidence of technical leadership (led a team or project); (4) industry experience (preferred, not required).

EBA process: Apply criterion 1 — reduces to 80 candidates. Apply criterion 2 — reduces to 35. Apply criterion 3 — reduces to 12. The remaining 12 are evaluated holistically, with criterion 4 as a differentiator.

Result: the hiring manager spent focused time evaluating 12 genuinely qualified candidates instead of reading 200 applications with variable attention. The EBA served its purpose — focusing limited evaluation resources on a manageable set — while the holistic final evaluation protected against the model's weakness (eliminating someone good on everything except the first criterion).

Apartment Search

Someone relocating to a new city has 50 apartments to evaluate. Their criteria: (1) maximum monthly rent of $2,500; (2) maximum commute of 30 minutes by transit; (3) minimum 2 bedrooms; (4) pet-friendly.

EBA: Apply $2,500 max → 30 remain. Apply 30-minute commute → 18 remain. Apply 2+ bedrooms → 11 remain. Apply pet-friendly → 6 remain. The final 6 are visited in person and evaluated on less quantifiable dimensions: natural light, building quality, neighborhood feel.

The EBA turned a 50-apartment research project into a 6-apartment decision — appropriate given the constraints.

Venture Capital Deal Screening

A seed fund evaluates hundreds of companies per quarter. Their EBA criteria in order: (1) B2B software or fintech — fund mandate, non-negotiable; (2) at least one technical co-founder — thesis requirement; (3) some form of early traction (customers, LOIs, or compelling pilot data); (4) raising a round the fund can lead ($1–3M check size range).

Applying these in sequence reduces a weekly flow of 40–60 opportunities to 3–5 that merit a full partners meeting. The EBA doesn't select the best companies — it efficiently eliminates the ones that are clearly outside the fund's mandate so partners can spend their analytical energy on the real decisions.

When to Use It

✅ Use Elimination by Aspects when:

You have many options (more than 5–7) and multi-attribute evaluation of all of them is impractical
One or two criteria clearly dominate and others are secondary
Time is limited and you need to narrow a candidate set quickly
The decision has clear minimum requirements (budget limits, compliance requirements, physical constraints)

❌ Use cautiously or supplement when:

All criteria are roughly equally important (EBA's sequential structure is less appropriate)
The option that fails criterion 1 by a small margin might be the best overall on criteria 2–10
The decision is high-stakes and irreversible — supplement with holistic evaluation of finalists
Criteria thresholds are uncertain — if you're not sure where to set the threshold, EBA can be highly sensitive to that uncertainty

Pairs well with	Why
Decision Matrix	Use EBA to narrow, then Decision Matrix to compare finalists
Satisficing	Both are "good enough" approaches that avoid exhaustive optimization
MECE thinking	Ensures your criteria set is complete before ranking
Opportunity Cost	Helps rank which criterion should be applied first

Common Misuses and Limitations

Applying criteria in the wrong order. EBA is highly sensitive to the order in which criteria are applied. If you apply a secondary criterion first, you may eliminate your best option before the primary criterion is even considered. Always apply the most important criterion first.

Setting thresholds too strictly. An option that's excellent on all criteria except criterion 1 — where it's just below the threshold — may be better than every option that passes. Strict binary thresholds prevent the model from capturing near-misses. Use EBA to narrow, then evaluate finalists holistically.

Using EBA for decisions with only a few options. If you have 4 options and 3 criteria, just evaluate them all. EBA's value is in managing cognitive load — when the candidate set is large. For small sets, use a decision matrix.

Not revisiting the criteria order. The order of criteria should reflect your actual priorities, which can change. Re-examine the ranking before applying EBA rather than treating a previous ranking as permanent.

Satisficing — both approaches prioritize "good enough" decisions over exhaustive optimization
Decision Matrix — the complement: use EBA to narrow to finalists, Decision Matrix to compare them
10-10-10 — helps clarify which criteria matter most over different time horizons

FAQ

How is EBA different from just filtering?

They're related — EBA is filtering done in a deliberate, criterion-ranked sequence. What distinguishes EBA is the explicit ranking of criteria by importance and the sequential (not simultaneous) application. Most informal filtering applies criteria in whatever order comes to mind. EBA applies them in order of importance, which ensures that the most important criteria do the most work.

What do I do when no options survive the elimination process?

Either the thresholds are too strict, the option set is too narrow, or your constraints are genuinely incompatible. First, identify which criterion eliminated the most options and ask whether the threshold is right. Then consider whether you can expand the option set. If not, you may be facing a genuinely constrained decision that requires compromising on one of your criteria — which is useful information.

Is EBA the same as a knockout criteria system?

Yes, essentially. Knockout criteria (or "eliminators") are criteria applied before more nuanced evaluation — any option that fails a knockout is removed from consideration. EBA is a formalized version of this that (1) ranks the knockouts by importance and (2) applies them sequentially rather than simultaneously, giving you structured control over how options are filtered.

Apply with AI

🚀 Run an Elimination by Aspects analysis in MindMax →

What Is Elimination by Aspects?​

How It Works​

Three Real-World Examples​

Hiring: Filtering a Candidate Pool​

Apartment Search​

Venture Capital Deal Screening​

When to Use It​

Common Misuses and Limitations​

Related Models​

FAQ​

Apply with AI​

Further Reading​