Reply taxonomies: coding responses for learning

Two reps run the same sequence for a month. Each gets about forty replies, and thirty of those are some flavor of no. The first rep logs them all as "not interested" and moves on. The second tags every reply with one of five codes before archiving it. At the end of the month the first rep has a discouraging number. The second has a map of the market, drawn by the market itself.

Reply rates on cold email have fallen into the low single digits, a slide we traced in the reply-rate numbers back in March. When replies are that scarce, wasting the information inside one is the expensive habit. A no is not a verdict. It is a data point with a category, and the category is what steers.

The five codes

Not now. The problem is real, the timing is wrong. "Circle back in Q3" is not a rejection, it is a scheduling instruction. The steer: a recycle date on the account and a light-touch hold pattern until it arrives. In a healthy program this is the biggest bucket, because timing is the one variable no amount of targeting controls.

Wrong person. A routing error, and often a gift, because half of these replies name the right person. The steer is upstream: if a third of your replies are wrong-person, your title targeting is off, and the fix belongs in the list build, not the message.

Already covered. An incumbent exists. This is quietly good news, because it confirms the account pays for the problem. The steer: find the renewal window, note the incumbent's category, and treat the account like the warm file we described in the closed-lost list. Vendors change. Budgets survive.

No need. The only reply that actually challenges your targeting thesis. If accounts that match your ICP on paper keep saying the problem does not exist for them, the ICP is wrong on paper too. This bucket should be small. When it grows, stop and re-examine the list before touching the message.

Never. The hard no. Suppress the contact, respect the answer, move on. It should be the smallest bucket by far. If it is growing, the problem is usually tone or frequency, and it is worth hearing that early.

Read the mix, not the count

The total number of nos tells you almost nothing. The proportions tell you nearly everything. A pile of wrong-person replies is a list problem. A pile of no-need replies is a targeting thesis problem. A pile of not-nows is, counterintuitively, a sign to hold steady: the market has confirmed the problem and merely disputed the calendar, and the patient cadence wins that dispute. The codes turn "outbound is not working" into a specific, fixable sentence, which is the difference between steering and sulking.

None of this works if the tags live in the rep's memory. Code the reply the moment it arrives, in the same field, from the same picklist, every time. Free-text notes rot; categories compound, which is the argument we made in notes fields are data. Five codes is few enough that reps actually use them and enough that the mix means something.

Every CommandVA rep codes every reply and every call disposition this way, and the weekly report your team gets includes what the mix moved and what we are adjusting because of it. If your current program hands you a reply count with no anatomy, book a strategy call and we will show you what a coded month looks like against an ICP like yours.