Lead Generation KPIs: The Six That Survive a Quarter

Most lead generation dashboards report the same six things, and five of them move without anything changing. Open rates drift with a mailbox provider's image-loading behaviour. Click rates drift with whether a link was in the signature. Impressions drift with somebody else's ad budget. A quarterly review built on those numbers produces a discussion about the numbers rather than about the pipeline.

A useful KPI has one property that most of the reported ones lack: when it moves, something you did caused it, and you can say which thing. That is a demanding test and it removes most of a standard dashboard.

The chain a lead generation KPI has to sit on

Before choosing metrics, it helps to name the chain they measure, because a metric that is not attached to a step in it is decoration.

Outbound has four transitions and each one is a place where volume is lost. A defined target set becomes a reachable list. A reachable list becomes contacted companies. Contacted companies become conversations. Conversations become meetings that happen and that meet an agreed standard.

Every legitimate lead generation KPI is either the size of one of those stages or the conversion between two of them. That framing does two useful things at once. It makes it obvious when a metric belongs to nobody, and it makes it obvious that improving a rate at one stage while shrinking the stage before it is not an improvement at all.

Defined target set

Companies matching written criteria a stranger could re-run

Reachable list

Those with a verified way to contact the right person

Contacted

Companies that received a message, counted once

Conversations

Replies that continue, positive or otherwise

Qualified meetings

Meetings that happened and met the agreed standard

The six worth keeping

Reachable rate. The share of your defined target set for which you can actually contact the right person. This is the earliest KPI and the most ignored, because it belongs to data rather than to messaging. A campaign against a market you can only half reach has a ceiling nobody wrote down, and no amount of copy work raises it.

Contacted coverage against the addressable set. How much of the market you have used. This is the metric that turns a good month into an honest forecast, because a list that has already been contacted is not available again on the same angle. Teams that do not track it plan the next quarter against a market they already spent.

Reply rate, read as a diagnostic rather than a score. Useful because it responds to things you control, and only useful when it is split. Reply rate by segment tells you where the offer lands. Reply rate by list source tells you whether the data is the problem. A single blended figure tells you almost nothing, which is why it is the number most often reported and least often acted on.

Positive-reply rate. The share of replies that are interest rather than refusal or noise. This separates a message that provokes from a message that persuades, and the gap between reply rate and positive-reply rate is one of the most diagnostic pairs on the list.

Qualified meeting count, against a written standard. The output metric, and the only one that survives a conversation with a finance team. It is meaningful only where the standard was agreed in writing before launch, which is why a qualified appointment is a definition rather than a feeling. A meeting count without a definition behind it is a number both parties can argue about after the fact, which is exactly when the argument is least productive.

Cost per qualified meeting. Total spend, including tooling, data and people, divided by meetings that met the standard. This is the number that lets outbound be compared with anything else you could have done with the money, and it is the one that makes a cheaper channel with worse conversion look correctly bad.

Notice what is absent. Open rate is not on the list, because it measures a mailbox provider's behaviour as much as a recipient's. Total leads is not on it, because a lead with no definition is a row. Activity counts are not on it, because they measure effort, and effort is an input a manager can already see.

Two properties that make a KPI usable

The metrics above share two structural properties, and those properties are more transferable than the specific list.

A denominator you chose deliberately. Reply rate over messages sent and reply rate over companies contacted are different numbers, and a team that switches between them without saying so has produced an improvement out of arithmetic. Write the denominator next to the metric, permanently.

A companion that moves the other way. Almost every rate can be improved by shrinking its denominator, so a rate reported alone invites exactly that. Reply rate pairs with contacted volume. Positive-reply rate pairs with reply rate. Qualified meeting count pairs with cost per qualified meeting. A pair is honest in a way that a single figure is not, and a dashboard of paired metrics is roughly half the size of a normal one.

The illustrative arithmetic in this paragraph is invented and describes no real campaign. A team contacts 2,000 companies and books 20 qualified meetings, so one company in a hundred becomes a meeting. The next quarter it contacts 500 of the best-fit companies and books 10, and the rate has doubled while the meeting count halved. Both quarters are correctly reported and only the pair shows what happened.

Before adding a metric

• ✓Yes: When it moves, can you name the thing you did that moved it.
• ✓Yes: Is the denominator written down beside it and stable across quarters.
• ✓Yes: Does it have a companion metric that shrinking the denominator would damage.
• ✓Yes: Does one named person own the number rather than a function.
• ✓Yes: Does it correspond to a stage or a transition in the funnel, rather than to activity.
• ~Depends: Would it survive being shown to a finance team without a paragraph of explanation.
• ✓Yes: Is it measured against a definition agreed before launch rather than after results arrived.

Who reads which number, and how often

A metric with no reader is a metric nobody maintains, and most dashboards die of that rather than of being wrong. Splitting the six by audience and by frequency solves more than refining the definitions does.

Reachable rate and contacted coverage are planning numbers. They are read before a quarter rather than during one, because they set what is possible, and reading them weekly produces no decision anyone can act on.

Segmented reply rate and positive-reply rate are operating numbers. They are read while campaigns are live, they are the pair that tells you whether to change the angle or change the list, and they are the only two on the list that reward frequent attention.

Qualified meeting count and cost per qualified meeting are accountability numbers. They are read at the end of a period, by whoever owns the budget, and they are the two that have to be defined before the period starts rather than after.

Assigning each number a reader also settles a quieter question, which is who is allowed to change a definition. The answer should be that nobody changes one mid-period, because a definition that moves makes every comparison across the boundary meaningless, including the comparison against your own last quarter that the section below argues is the only reliable one.

Where benchmarks belong, and where they do not

The first question after any KPI discussion is what good looks like, and it is the question most likely to send a team in the wrong direction.

External benchmarks are worth reading as orientation and they cannot settle a target, because published figures average across industries, list qualities, offers and definitions of the metric itself. Two companies reporting the same reply rate can be counting different events. Where a comparison is useful, the cold email benchmark set collects published figures with their sources attached, which is the right shape for orientation.

The comparison that actually decides anything is against yourself. Your own last quarter, measured the same way, with the same denominator, is a control that no external figure can be. It holds constant the offer, the market and the definition, which are the three variables a benchmark cannot.

That is also the honest answer to the target-setting question. A target derived from your own trajectory is a plan. A target derived from someone else's median is a wish with a citation.

Attribution, and how much of it to attempt

One structural warning belongs here because it consumes more analyst time than any other question in this space.

Multi-touch attribution in outbound is expensive to build and rarely changes a decision. The useful version is much smaller: record which list source and which angle produced each qualified meeting, and stop there. Those two fields answer the questions a plan actually asks, which are what to build more of and what to stop doing.

Everything past that is an argument about credit allocation between channels, and it is worth having only when the answer would change a budget. It usually would not, because the constraint is nearly always the size of the reachable set rather than the split of credit for the meetings you already got.

The related decision, and the one that changes cost more than any dashboard, is whether the motion is bought or built. Appointment setting against lead generation covers what is being purchased in each case, and b2b lead generation services covers how the work is scoped when it is bought. Both start from cost per qualified meeting, which is why that metric sits at the bottom of the list above.

What to take away

Keep six metrics: reachable rate, contacted coverage against the addressable set, segmented reply rate, positive-reply rate, qualified meeting count against a written standard, and cost per qualified meeting. Drop open rate, total leads and activity counts, which move without anything changing.

Write the denominator beside every metric and give every rate a companion that shrinking the denominator would damage, because a rate reported alone can always be improved by doing less.

Use external benchmarks for orientation and your own previous quarter as the control, since it is the only comparison that holds the offer, the market and the definitions constant. And agree what counts as a qualified meeting in writing before launch, because a definition settled after the results arrive is not a definition.

Where the constraint is the reachable set rather than the reporting, RevenueFlow books qualified meetings on a pay-per-meeting basis, against criteria agreed in writing before launch, and what a lead costs to generate is the arithmetic that sits underneath the comparison.