Employee Ideas to Experiments

Start an Improvement Turn a work problem into a small experiment For employees who notice a concrete problem and want to test a practical improvement. This fictional example follows Northstar Supply, a multi-site organization. Use the pages in this site to move from an observable problem to a bounded test and shared learning. Choose your next step Start where you are; each page includes a practical example. Frame the problem Describe what happens before deciding what to change. Shape an idea Make the question, scope, and evidence reviewable. Run a small experiment Agree a bounded test with the relevant work owner. Share results Report observations, limitations, and a next-step recommendation. A sample result, not a live metric Illustrative Northstar Supply example. These figures are fictional and are not evidence about your organization. Before: handoffs with next owner and open item recorded 6 of 10 During test: handoffs with next owner and open item recorded 8 of 10 How to use the sample figures Illustrative source: a fictional manual count of 10 handoffs in each of two short test periods. The unit is handoffs where the next owner and open item were recorded. The example comparison is 6/10 before and 8/10 during the test; the fictional as-of date is 14 May 2025. This small example shows how to name a measure, numerator, denominator, source, and date. It does not establish cause, represent all sites, or describe real Northstar operations. Frame the Problem Describe what happens before proposing a fix A useful frame names the place or task, an observable pattern, who or what is affected, and what evidence would clarify it. Use this sentence pattern: During [task or situation], we observed [specific event or count]. This affects [work, handoff, or group]. We need to learn [open question] by checking [observable evidence]. Keep the description neutral. Separate what you saw from what you think caused it, and avoid including personal or sensitive information that is not needed to understand the work. Make the problem observable Fictional Northstar Supply example: a shift handoff sometimes leaves the next team unsure which open item needs attention. Too broad to test “Shift handoffs are bad and people need to communicate better.” Does not identify where or when the issue occurs. Blames people without describing an observable event. Does not say what evidence could clarify the problem. Specific enough to explore “In the evening-to-morning handoff, the next owner is not recorded for some open requests. We want to count how often the next owner and open item are both clear at the start of the following shift.” Names the handoff and the missing information. Describes a checkable observation. Leaves the cause and best change open for discussion. Gather enough context to frame it Use observations that are relevant to the work question, not a large data collection exercise. Pin down the moment Name the task, location or handoff, and the situation in which the issue appears. Record a small set of observations For the sample handoff question, note whether both the next owner and open item are recorded. Keep the unit and observation period clear. Check whether others recognize the pattern Ask people who perform or receive the work what they observe. Note differences by shift or site rather than assuming one account applies everywhere. Name boundaries and dependencies Identify the work owner and any approved procedure, system, or other team that constrains possible changes. Do not use an informal experiment to replace a controlled process. Shape an Idea Make an idea reviewable A useful proposal connects a framed problem to a test question, a bounded change, observable evidence, and a responsible work owner. State what you want to learn, not only what you want to install or change. Describe who and what the test covers, what will be compared, and what could make the observation misleading. The relevant work owner can help decide whether the proposed scope is appropriate. Completed sample idea record Fictional Northstar Supply shift-handoff example. The record is an illustration, not a submitted or approved idea. Record field Illustrative entry Problem Some evening-to-morning handoffs do not clearly record the next owner and open item. Proposed change Try a short handoff prompt asking for the open item and its next owner. Test question During the agreed test period, are both details recorded in more handoffs than in the starting observations? Owner role The work owner for the handoff reviews scope and helps interpret the result. Scope A small, agreed sample of one handoff context; do not assume findings apply to other shifts or sites. Evidence to compare Count handoffs where both details are recorded; state the count, denominator, source, and observation dates. From broad suggestion to testable proposal The change is not the question. Make the learning goal explicit so the work owner can assess the test. Initial suggestion “We should replace the handoff process with a new system.” The change is large before the problem is understood. The scope, dependencies, and measure are unclear. It may affect procedures or tools beyond the proposer’s role. Reviewable first test “Could we try a brief prompt for recording the next owner and open item in an agreed handoff context, then compare a small count with the starting observations?” Names a limited change and a learning question. Leaves scope and duration to agreement with the work owner. Makes a result possible to discuss without claiming proof. Run a Small Experiment Keep the test within agreed work boundaries Test only changes that are within your role and agreed with the relevant work owner. This improvement process does not replace approved safety, compliance, security, or other controlled procedures. If a proposed change could affect one of those areas, pause and use the applicable approved route with the responsible trained role. Plan and carry out a bounded test The example sequence is for work coordination; agree the actual scope with the work owner. Agree the question and scope Confirm what the test is meant to learn, which work context it covers, and who owns the relevant process. Record a starting observation Choose an observable measure and record its unit, count, denominator, source, and dates. Use only information appropriate for the work question. Agree the change and observation period The work owner and proposer agree what will change, where it applies, who will do it, and when the observation ends. Carry out only the agreed change Stay within the agreed scope. Stop and consult the work owner if the change creates an unexpected impact or crosses a controlled-process boundary. Capture what happened Record the same measure where practical, note deviations and limitations, and bring the result to the work owner for discussion. Compact experiment record Illustrative completed example on the right. It is not a status report, approval, or operational instruction. Field Fictional example Question Did the prompt help record both the next owner and open item? Scope One agreed handoff context; findings are not assumed to apply elsewhere. Starting point 6 of 10 sample handoffs had both details recorded; fictional manual count, dated 7 May 2025. Change Try a brief prompt for the open item and next owner, as agreed with the work owner. Observation period A short fictional sample period ending 14 May 2025; an actual period must be agreed for the real work. Result 8 of 10 sample handoffs had both details recorded in the later fictional count. Limitations Small illustrative counts; no control group; short period; no evidence about other sites or causes. Next decision Discuss with the work owner whether to adjust, gather more information, or stop this test. Who owns the handoff? The employee proposing the test describes the problem and records observations within the agreed scope. The relevant work owner decides whether a change can be explored in that work context, helps set boundaries, and discusses what the evidence can and cannot show. If the issue touches a controlled process, the responsible trained role and the applicable approved procedure take precedence. Share Results Share what happened, not just what you hoped Useful learning includes results that are mixed, inconclusive, or different from expectations. This project-work image is illustrative stock photography, not a depiction of Northstar Supply or its employees. A concise results structure Give the reader enough detail to judge the observation and its limits. Question: What did the test try to learn? What was tried: What change was agreed, and within what scope? What was observed: Give the measure, unit, count, denominator, source, and dates. Limitations: What was too small, different, incomplete, or uncertain to conclude? Recommendation: Should the team continue discussing the test, adjust it, stop, or gather more information? Separate an observation (“8 of 10 sample handoffs recorded both details”) from an interpretation (“the prompt may have helped in this context”). A small before-and-after count does not establish why a result changed. Worked example: the shift-handoff prompt Fictional results, offered as a model for careful reporting. Question and change Would a brief prompt help record the next owner and open item? In the fictional test, a prompt was tried in one agreed handoff context. Starting observation: 6 of 10 sample handoffs recorded both details. Later observation: 8 of 10 sample handoffs recorded both details. Source: fictional manual count; sample dates 7 and 14 May 2025. Interpretation and limit The later sample contains two more handoffs with both details recorded. The example does not show that the prompt caused the difference or that it would work in another context. The samples are small and do not include a comparison group. Other conditions may have changed between observations. The counts describe this fictional example only. Next discussion The proposer shares the record with the relevant work owner. Together they can discuss whether an adjusted test or more information would be useful; the figures do not automatically approve a wider change. Name who should review the result. State what decision or question is next. Keep the recommendation proportionate to the evidence. Choose a proportionate next discussion These are discussion options, not automatic approvals or a substitute for the work owner’s decision. What the evidence looks like A reasonable next discussion The measure moved in the intended direction, and the scope stayed as agreed. Discuss whether a further bounded test would answer the remaining question. The result is mixed, or the prompt was difficult to use consistently. Adjust the question or change with the work owner, then decide whether another observation is useful. The change created an unwanted effect or crossed an agreed boundary. Stop the test and raise the issue with the relevant work owner; follow applicable approved procedures. The sample is too small, inconsistent, or incomplete to interpret. Gather more appropriate information or report that no clear conclusion can be drawn.