operations

The Disposition List Is the Only Data You Get

Every call produces exactly one fact, chosen by a tired person from a dropdown. The design of that dropdown decides what the business can know.

A call ends and an agent picks a disposition. That single choice is the entirety of what the business will ever know about the call.

Most disposition lists are designed by whoever set up the dialler, in ten minutes, and never revised. The consequences run for years.

The two failure shapes

Too few options. A list of five forces unlike outcomes into one bucket. "Not interested" absorbs a person who already has a contractor, a person who is not the decision maker, a person who was called at a bad moment, and a person who is never going to buy. Those four need different follow-up and the data cannot tell them apart.

Too many options. A list of forty is slower than the call it describes. Agents respond by learning three of them and using those for everything, which produces data that looks detailed and is not.

The useful range is around eight to twelve, grouped, with the most common three at the top.

What every list needs

Four categories, and most lists are missing the second and third.

Contact outcomes. Reached the right person and something happened: interested, appointment set, quoted, not interested.

Reachability outcomes. No answer, voicemail, busy, disconnected, wrong number. These are facts about the number, not about the prospect, and conflating them with the first group is what makes list quality invisible.

Routing outcomes. Not the decision maker, call back later with a date, asked for information first. These are not failures and most lists record them as one, which means the best outcome available on a cold call, being told who to speak to, is logged as a loss.

Compliance outcomes. Asked to be removed, asked not to be called again, requested no recording. These are not sales outcomes at all. They are obligations with a date, and they must write to suppression rather than to a report.

That last category is the one with legal weight. An internal do-not-call request under the Telemarketing Sales Rule at 16 CFR part 310 has to be honoured, which means the disposition has to actually do something rather than being a label on a row.

Make the consequence automatic

A disposition should change the state of the record, not just describe it.

"Call back Tuesday" should create the callback. "Wrong number" should flag the number rather than the person. "Do not call" should suppress across every campaign, immediately, including ones created later. "Appointment set" should create the appointment.

Where the disposition and the consequence are separate steps, the second one is skipped under pressure, and the gap between what agents selected and what the system did is invisible until somebody is called who asked not to be.

One required field, at most

Some dispositions need more: a callback needs a date, a removal request needs no detail at all.

The temptation is to require a note on every call. Required free-text fields on high-volume work produce "n/a" at scale, which is worse than nothing because it looks like data.

Require a structured field where it changes behaviour, which is the callback date. Make notes optional and they will be written when there is something to write.

Audit the distribution

Two reports, quarterly, and they take ten minutes.

Disposition distribution by agent. Outliers are nearly always a training issue rather than a performance one: an agent using "not interested" for everything is not worse at calling, they are using a different definition.

Share of calls in the top three dispositions. If it is over ninety percent, the rest of the list is decorative and the data is coarser than it appears.

The connection to everything else

The disposition list is upstream of every number the outbound operation reports. Connect rate depends on separating reachability from contact. Answer-rate analysis depends on knowing which numbers are dead. Compliance depends on removal requests writing through. List quality depends on wrong numbers being distinguishable from no-answers.

Ten minutes designing the dropdown is worth more than any dashboard built on top of it, and almost nobody spends them.