The short version
Early on, I measured my own effort in activity: calls made, emails sent, accounts researched, because activity was controllable and outcomes felt like luck. The shift that actually moved results was measuring quality of judgment instead: was this the right account to prioritise, was this message built around their actual trigger, was this the right moment to pick up the phone versus send an email.
Activity without judgment produces busy weeks and flat pipelines. The uncomfortable lesson was that doing less, more precisely, consistently beat doing more, generically, and that precision is a skill you build by reviewing your own misses, not by working longer hours.
Defining a quality of judgment metric
Quality of judgment sounds abstract until you turn it into a checklist. For each touch I ask four yes-or-no questions: Is the account in the top tier by fit? Is there a verified trigger from the last 90 days? Is the contact at the right level for the problem? Is the channel suited to this person? A touch scoring four of four is a high-judgment touch; track the percentage of those.
Run the checklist before sending, not after. It takes about 20 seconds and filters out roughly the weakest tenth of what you were about to do. Over a month, compare reply rate for four-of-four touches against the rest. If the gap is small, your checklist is wrong, and that is useful information too.
Building a miss log that you actually read
A miss log needs a fixed format or it becomes diary writing. Use five columns: date, account, what I did, what the signal was, what I would do now. Add one entry per week of something that failed in a way you could have predicted, such as a call to a person who had left the company or a message that cited the wrong product line.
Then code each miss: wrong account, wrong person, wrong trigger, wrong timing, wrong message. After eight to ten entries a single code usually dominates. That is your improvement target for the next month. It is more effective than a general resolution to be better, because it points to one specific habit to change.
Finding enough volume to learn from
Precision has limits: you need sample size to see patterns. A rough rule is at least 30 to 50 touches per variable before you trust a comparison. If you work 15 priority accounts per week, test one change per fortnight, not three at once. Otherwise you cannot tell which change moved the result.
Use tiers to protect learning. Keep a small experimental tier, perhaps 20 percent of your touches, where you try different angles on lower-priority accounts. Reserve your top tier for the best-known approach. This way you learn at low cost and apply the winners where the stakes are high, which avoids the trap of either blind volume or timid perfectionism.
Worked example (illustrative)
Hypothetical: a researcher cuts daily outbound touches from 60 to 35 and adds a four-question pre-send checklist for one quarter.
What changed: Fewer, better-judged touches produced almost double the meetings. The driver was the trigger requirement, which filtered out accounts with no reason to reply now.
Do it: the steps
Tick them off as you go. Your progress stays on this device.
Traps to avoid
- ×
Overcorrecting into doing so little that there is nothing to learn from. Precision needs enough volume to generate evidence.
- ×
Measuring judgment by gut feel. Without a checklist and a log, you will remember hits and forget misses, which makes every approach seem to work.
- ×
Changing the target list and the message at once. If results move, you will not know which change caused it, so alter one variable per test window.
What to track
- Percentage of touches with a verified trigger
- touches with a logged trigger divided by total touches; aim for over 60 percent on tier-one accounts.
- Meetings per 100 touches
- a normalised view that rewards precision; typical outbound ranges are 1 to 5, depending on the segment.
- Miss-log completion rate
- weeks with at least one logged miss divided by weeks worked; target 100 percent.
Benchmarks are typical planning ranges, not guarantees. Calibrate to your own data.
Questions people ask
How do I know whether I am being precise or just slow?
Compare output per hour, not per day. If your meetings per hour of work rise while touches fall, you are precise. If both fall, you are probably over-researching. Cap research at a fixed time per account, for example eight minutes for tier one, and stop when the clock ends.
What is a verified trigger?
A dated, public event that creates a reason to talk now: a funding announcement, a new executive, a job posting in the relevant function, a product launch or regulatory change. Verified means you saw the source yourself within the last 90 days, not a vendor’s intent flag you cannot explain.