Quick answer
An evidence log records the decision, observed facts, source, date, interpretation, uncertainty, and next test. Keep facts separate from conclusions, use one row per claim or signal, and revisit the log after the opportunity or project produces new information.
Independent professionals make decisions with incomplete information. A lead sounds promising but has no clear decision owner. A buyer requests a familiar service but cannot describe the outcome. A project looks profitable until access delays and review cycles consume the capacity you reserved. If you keep only the final conclusion, you lose the reasoning that could improve the next decision.
An evidence log is a simple record of what you knew, what you inferred, what remained unknown, and what you did next. It is not a diary and it does not need to become a complex CRM. Its purpose is to make judgment inspectable.
Start with the decision
Write the decision at the top of the record. Examples include:
Should I pursue this opportunity?
Is this service worth testing with this buyer group?
Should I accept the proposed scope and fee?
What evidence would justify changing my offer?
Should I renew, expand, refer, or offboard this client?
The decision gives the evidence a job. Without it, the log becomes a collection of interesting facts that do not improve action.
Use six fields for each signal
Each row should contain:
Observed fact: what happened or what the source states.
Source: where the information came from.
Date: when the fact was observed or published.
Interpretation: what you think the fact may mean.
Uncertainty: what could make the interpretation wrong.
Next test: the smallest action that could clarify the issue.
For a lead, the observed fact might be that the buyer supplied a deadline, named a decision maker, or asked for a specific output. The interpretation might be that the project is ready for a scoped proposal. The uncertainty could be whether budget is approved. The next test could be a focused question about approval and access.
Do not put “good lead” in the fact field. That is an interpretation. The distinction may feel fussy, but it prevents confidence from disguising missing evidence.
Capture source and scope precisely
Source quality depends on what was actually observed. “Client call” is better than no source, but “September 29 call with operations director; client said the reporting process takes three days each month” is more useful. For public research, retain the publisher, final URL, publication date, metric definition, population, and limitation. For your own opportunity data, record the source channel, sample, and time window.
This discipline matters when you compare patterns. Ten referral conversations are not the same evidence as ten marketplace listings. A client’s estimate of a problem is not the same as a measured baseline. A public occupation projection is not a live count of contract demand.
Record disconfirming evidence
Most notes preserve reasons to proceed. Add a deliberate field for evidence against the current interpretation. Maybe the buyer likes the concept but will not provide access. Maybe several prospects use the same language but none can name a budget or decision date. Maybe the project is profitable on paper but repeatedly displaces higher-fit work.
Disconfirming evidence does not automatically mean “stop.” It tells you which assumption deserves attention. You may narrow the scope, change the buyer segment, adjust the proof, or decline the opportunity.
Use confidence labels carefully
A simple confidence label can help if it is tied to evidence. Use low when the conclusion rests on one unverified statement, medium when multiple relevant signals agree, and high only when the decision has been tested across a meaningful sample or outcome. Do not let a high-confidence label mean “I feel strongly.”
You can also label evidence class: public research, SmartBid-observed opportunity data, client-provided information, personal capacity constraint, or interpretation. Keeping classes separate reduces accidental overclaiming.
Review at three moments
Review the log when you make the decision, when the work starts, and when the outcome is known. At the first review, ask whether the next action is proportionate to the evidence. At kickoff, check which assumptions became contractual or operational dependencies. After the outcome, record what happened and which signal predicted it or failed to predict it.
The outcome review should not rewrite history. Keep the original interpretation visible and add the new information below it. Learning requires knowing what you believed before the result.
Connect the log to your operating system
An evidence log becomes useful when it changes a recurring process. If many opportunities fail because the decision owner is unclear, add a qualification question. If proposals are delayed by missing access, name access as a precondition. If a service attracts interest but no concrete next step, rewrite the offer around a decision and deliverable.
Use the log to update:
opportunity qualification;
discovery-call questions;
proposal assumptions and exclusions;
proof and case-study selection;
capacity planning;
pricing floors and payment terms; and
renewal or offboarding criteria.
Do not use it to create a false sense of control. Some decisions remain uncertain because the world is uncertain. The log helps you see the uncertainty and choose a sensible next action.
A worked example
Decision: Should I pursue a six-week operations project for a growing services firm?
Observed facts: The buyer described repeated reporting delays, named the operations lead as the sponsor, requested a six-week start, and has not confirmed access to the source systems. Source: discovery call and follow-up email, dated September 29.
Interpretation: The problem may fit a reporting-process diagnostic, but the implementation scope is not yet defined. Uncertainty: access, decision authority, and whether the sponsor controls budget. Next test: send a bounded diagnostic option with an access checklist and a decision meeting, rather than quoting the full implementation.
The record supports a next step without pretending the project is already qualified. If access is confirmed, the opportunity improves. If access is unavailable or the sponsor cannot approve the work, the record supports a decline or pause.
Methodology and limits
This is a non-quantitative decision framework. The examples are illustrative. An evidence log does not replace contractual review, professional judgment, legal advice, financial advice, or source-specific data quality checks. For quantitative market claims, retain an evidence packet with the source, metric definition, scope, time window, freshness, limitations, and reviewer.
SmartBid helps independent professionals evaluate opportunities, prepare for pursuit decisions, and track outcomes. The log remains yours; its purpose is to make your reasoning more useful the next time a similar decision appears.