01 · Definition and the local convention
Definition and the local convention
The N/3 Method here is a local common multivoting convention: each participant receives approximately one third as many votes as there are ideas on the eligible list. It is a practical way to narrow a long collection of proposals. The convention must state how fractional results are rounded, whether a participant may place multiple votes on one idea, and what happens at a tie. Different organizations use different rules.
This lesson uses N as the number of ideas after duplicates and ineligible concepts have been removed. Each participant receives floor(N/3) votes, with a minimum of one when the list contains fewer than three ideas. Each vote goes to a different idea. These are declared local rules, not a mathematical truth or proof of solution quality. The output is a preference shortlist that still needs evidence, feasibility assessment, and explicit treatment of risk.
02 · When voting is useful and when it is not
When voting is useful and when it is not
Use this convention when a group has generated many reasonable ideas and needs to choose a manageable number for further evaluation. It works best when participants understand the proposals and have broadly comparable access to information. It is poorly suited to deciding whether a safety requirement can be ignored or whether a cause has been demonstrated. Popularity cannot replace a must criterion or a causal investigation.
Fictional Northbank Dispatch has eighteen eligible dispatch improvement ideas after combining duplicates. They include lane markings, early carrier assignment, scan confirmation, printer relocation, standard reprint handling, and revised order-release timing. The facilitator checks that none is merely a vague aspiration such as “better communication.” Each proposal has a short explanation and its known constraints. Participants may ask clarification questions before voting, but lobbying during ballot collection is avoided so quieter staff can express independent preferences.
03 · Run the method reproducibly
Run the method reproducibly
With N = 18, floor(18/3) gives six votes per person. Eight participants therefore have forty-eight possible votes. Every participant marks six different ideas on a private ballot. The facilitator explains that they are choosing ideas worth examining, not authorizing implementation. Private voting reduces social pressure, especially where supervisors and operators sit together. Names are retained only if the organization needs an audit; the public result does not reveal individual choices.
The group agrees in advance to shortlist the top four ideas, including any tie at fourth place. It also permits a separate risk flag for an idea that is essential to traceability or safe recovery. Risk flags do not add preference votes. This separation matters: combining risk concerns into the vote total can make a mandatory control look optional. After tallying, participants check that the totals match the ballot rules before interpreting the result.
04 · Worked tally and shortlist
Worked tally and shortlist
Northbank’s results are: early assignment eight votes, scan confirmation seven, printer relocation six, standard reprint handling six, lane markings five, revised release timing four, and the remaining twelve ideas together twelve votes. The total is 8 + 7 + 6 + 6 + 5 + 4 + 12 = 48, matching eight participants with six votes each. No single idea can receive more than eight under the one-vote-per-person-per-idea rule.
The top four ideas proceed to detailed comparison. The six-vote tie occupies third and fourth places, so it does not expand the shortlist in this round. A separate manual fallback proposal receives only one preference vote but a traceability risk flag. The team retains it as a required control for relevant concepts. The result thus narrows optional investigation while preserving a control that popularity would otherwise hide. It does not claim that eight votes predicts twice the benefit of four votes.
05 · What the tally means
What the tally means
A vote means a participant wants an idea examined under the stated question. It may reflect perceived benefit, familiarity, ease, or personal experience. Those reasons should be discussed after results are visible. The facilitator asks proponents and skeptics to explain the mechanisms and evidence, then captures disagreements for the next selection tool. A narrow difference between six and five votes is not statistical evidence that one option is superior.
If the team repeats the round with fewer ideas, it declares the new N and the new allowance. It does not quietly retain six votes after the list shrinks to seven. Repeated rounds can improve focus but also amplify an early majority and remove useful diversity. Northbank therefore stops after one narrowing round and applies must criteria, an effort benefit comparison, and a pilot design to the shortlist. Voting has done its bounded job.
06 · Limitations and advanced adaptation
Limitations and advanced adaptation
Weak use publishes a tally without the ballot rules, treats votes as effect measurements, or lets participants allocate all votes to a favorite while others assume distinct choices. Strong use defines N, rounding, vote placement, tie handling, and the decision question before collection. It checks arithmetic, preserves minority safety concerns, and documents why a shortlisted option is later accepted or rejected.
Advanced teams compare preference patterns by role, while avoiding unnecessary identification of individuals. If operators favor reprint handling and systems staff favor early assignment, the difference may reveal hidden assumptions about disruption or workload. A second anonymous discussion can resolve those assumptions before another ballot. The team can also reserve a small investigation slot for a low-vote but plausible novel concept. That is an explicit portfolio choice, not an attempt to manipulate the tally. The convention supports informed judgment when its limits remain visible.
07 · Operating conditions and evidence
Operating conditions and evidence
The facilitator checks whether the eligible list itself biases the vote. One concept described as “automated early assignment” may sound more developed than another described as “paper fallback,” even when both are equally necessary. Short descriptions should present mechanism, scope, and known limitation with comparable detail. Several variants of the same favorite can also split its votes or crowd out alternatives. Northbank consolidates genuine duplicates before defining N, while preserving materially different designs.
A tie rule matters even when the example does not require it. If lane markings had also received six votes, the agreed include-ties rule would produce five shortlisted ideas rather than four. The team would accept that result or use a separately declared follow-up comparison, rather than changing the rule after seeing the favorite at risk. This is procedural fairness with a practical purpose: participants can trust the narrowing process while understanding that its outcome is still subject to feasibility and evidence.
08 · Decision record and next step
Decision record and next step
Northbank publishes the shortlist with a note about what happens next. Early assignment and scan confirmation will receive a mechanism and feasibility review; the reprint concepts will receive a separate operational comparison. A vote does not create a budget commitment or oblige the team to implement every shortlisted proposal. Participants can see that the tally has consequences without confusing preference with authorization.
The team also preserves reasons for not advancing ideas. Some receive few votes because their benefit is unclear; others because they require another function. A future project can revisit those ideas when information or scope changes. The voting record should not turn temporary preference into permanent exclusion. A disciplined facilitator keeps the result bounded, checks the arithmetic, and carries essential minority concerns into the next decision. The method’s value is efficient narrowing with transparent rules, not an appearance of precision beyond what the ballots provide.
| Declared rule | Northbank choice | Why it matters |
|---|---|---|
| Definition of N | Eighteen distinct eligible ideas | The vote allowance uses the post-screening list |
| Rounding and placement | Floor N/3; one vote per chosen idea | Every participant follows the same convention |
| Shortlist and ties | Top four, with any tie at fourth included | The cutoff is known before results appear |
Connections
Related methods and phase use
Placed under Improve because of the outcome it supports. Use the method in other phases when its purpose and assumptions fit the work; the phase list is a guide to selection, not a compulsory sequence.
Sources
Professional foundations and further reading
The explanation and fictional teaching examples are written for LeanTrainingLab. Linked primary sources provide professional foundations; example values and local conventions are stated in the guide. The accepted toolset draws on Six Sigma+Lean Toolset, second edition (Lunau, editor; Springer, 2013). No book text or examples are reproduced here.