How to calculate contact rate, and the mistake most teams make

Contact rate = unique leads reached in a right-party, two-way conversation within a fixed window of arriving (14 days is a sensible default) ÷ unique leads in that same intake cohort. The mistake most teams make is dividing a calendar month of calls by the wrong denominator: in the worked example below, one September reports anything from 14.2% to 47.3%, against a true 41.0%.

How do I calculate contact rate? The formula

The formula that survives scrutiny has four parts, and each one closes a specific loophole:

Contact rate (D14) = unique leads created in the cohort that had a right-party, two-way conversation within 14 days of creation ÷ unique leads created in the cohort.

• Unique leads, not records. De-duplicate first. A person who fills in two forms is one lead, and counting them twice inflates the denominator and hides the fact that they were never called.
• Right-party, two-way. A conversation with the person who enquired, or the decision maker on a B2B list. Voicemails, ringouts, gatekeepers and delivered SMS are attempts. An SMS reply or an inbound callback is a contact.
• A fixed window per lead. Every lead gets the same 14 days to be reached, measured from its own creation time, so a lead that arrived on the 29th is not judged on two days of effort.
• A cohort, not a calendar. The numerator and denominator are the same people.

None of this is new. Survey researchers formalised it long before sales teams did: the American Association for Public Opinion Research’s Standard Definitions (revised 2023) publishes three formal contact rates, CON1 to CON3, built from final case dispositions rather than individual call attempts, and it says temporary, attempt-specific codes should be replaced with final disposition codes once each case’s final outcome is determined. Sales teams mostly skipped that step. Contact rate is a property of a lead, not of a dial, which is why it can only be calculated once each lead’s window has closed.

What is the mistake most teams make when they calculate contact rate?

They take a calendar month, count the conversations logged in that month, and divide by the leads created in that month. It feels right and is wrong in two directions at once.

It inflates the numerator. Conversations logged in September include August leads that were finally reached in September, plus repeat conversations with leads already reached. Neither belongs to September’s intake.

It deflates the numerator too. Leads created in the last two weeks of September have not had their full window yet, so the month closes before most of them could be reached. In survey terms these leads still carry temporary, not final, dispositions; analysts call it right-censoring.

Because the two errors pull in opposite directions, the monthly number can look stable while the underlying performance moves, and it swings sharply whenever lead volume rises or falls between months. The calendar-month contact rate measures when your team happened to talk to people, not whether this month’s leads were reached. We call the fix the matured-cohort rule: never report a contact rate for a cohort until its youngest lead has had the full window.

Worked example: one September, seven contact rates

An illustrative campaign, with round inputs so you can substitute your own. September produced 1,290 CRM records; removing 90 duplicates leaves 1,200 unique leads, of which 84 had invalid numbers and no other working channel. Across September the team logged 4,300 dials and 610 “connected” dispositions. Those connects covered 520 unique people, 70 of them August leads. Of the 450 September leads spoken to by 30 September, 40 conversations were with the wrong person. When the cohort matured on 14 October, 492 September leads had had a right-party conversation within 14 days of arriving.

The dashboard number overstates the true rate by 6.3 points; the right-party calendar number understates it by 6.8. A manager comparing either against last month is comparing two different things. The strict and working versions are both legitimate and mirror AAPOR’s CON1 (unknowns kept in the base) and CON3 (known-eligible only): report both, because the 3.1-point gap between them is your data-quality problem, measured. If that gap is large, the diagnostic tests for a low contact rate tell you whether the list or the calling is at fault.

Which contact rate should I use for which decision?

Each variant answers a different question. Using the wrong one is how teams end up buying a new list to fix a cadence problem.

In the example, 3,690 attempts went to the September cohort inside their windows, so attempts per contact is 3,690 ÷ 492 = 7.5. That single number converts a contact-rate target into caller-hours. A contact rate without its window and its denominator written next to it is not a metric, it is a mood.

How to split contact rate by attempt, channel and time of day

Once the cohort total is right, break it down by the attempt, channel and hour that produced each lead’s first right-party contact. Attributing only the first contact is what makes the slices add back up to the total, so you can see where your conversations actually come from.

• By attempt number. In the example, 170 first contacts came on attempt one, 118 on attempt two, 164 on attempts three to six and 40 on attempt seven or later. That is 492. More than two in five (204 of 492) arrived on attempt three or later, which a two-attempt cadence never makes.
• By channel. 331 answered outbound calls, 58 inbound callbacks, 81 SMS replies and 22 email replies, again 492. A callback or SMS reply is often the only path to the leads who screen unknown numbers.
• By hour band. Divide contacts by attempts within each band (for example before 12pm, 12–5pm, after 5pm, weekend). This one is a per-attempt rate, because the question is which hours convert a dial into a conversation.

The gap between a dialled attempt and a reached conversation is the whole story of this stage, and how to increase contact rate works through the attempt curve lever by lever.

How many leads do you need before a contact rate means anything?

A contact rate is a proportion, so it carries sampling noise. The approximate 95% margin of error is 1.96 × √(p × (1 − p) ÷ n). At a 40% contact rate:

Below about 200 leads a cohort, a month-on-month change in contact rate of less than 7 points is indistinguishable from noise. Smaller teams should use quarterly or rolling 90-day cohorts rather than chase monthly movements.

What does a higher contact rate do to sales, and what does measuring it cost?

In our own client work we typically see that doubling contact rate roughly doubles conversion from the same leads, and that speed to lead alone is worth around 3x. Those are our operating observations, not a published study, with no sample or window behind them, and not a guarantee. They also do not multiply: 3x and 2x is not 6x, because responding faster is one of the main ways contact rate rises in the first place. How we define and measure lift is on our methodology page.

The independent evidence on the related lever is separate and older. In “The Short Life of Online Sales Leads” (Harvard Business Review, 2011), Oldroyd, McElheran and Elkington analysed 1.25 million leads at 42 US companies and found firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it, defined as a meaningful conversation with a key decision maker, as firms that tried an hour later.

Measuring it properly costs less money than skill. You need three things: a de-duplicated lead ID, a “right party reached” disposition that reps actually use, and a first-contact timestamp per lead. The ongoing time is disposition auditing: checking 100 logged conversations a month at about three minutes each is five hours. What breaks at volume is not the spreadsheet but the discipline; dispositions drift as soon as nobody checks them. This stage feeds the next one, so the same cohort logic applies when you calculate appointment set rate, and the full chain is laid out in the sales pipeline stages and what each one costs. If you would rather hand the attempts and follow-up to someone else, that is what AI appointment setting covers.

Frequently asked questions

What is the difference between contact rate and connect rate?

Connect rate is connected calls divided by dial attempts, a per-dial number that tells you about list quality and caller ID health. Contact rate is unique leads reached in a right-party conversation divided by unique leads, a per-person number that tells you whether your leads were actually worked. In the worked example the same month gives a 14.2% connect rate and a 41.0% contact rate.

Should I include bad phone numbers in the contact rate denominator?

Report it both ways. Survey research does the same thing: the AAPOR Standard Definitions define CON1, which keeps cases of unknown eligibility in the base, and CON3, which includes only known-eligible cases. Use the strict version to judge the team and the working version to judge lead sources; the gap between them measures data quality.

How long should the contact window be?

Long enough for your full cadence to run, and identical for every lead. Fourteen days suits most inbound cadences of six or so attempts. Use a separate one-hour window to measure speed to lead, because a lead first reached on day ten counts in the 14-day rate but not in the one-hour rate.

Does speed to lead change contact rate?

Yes, and it is the best-evidenced lever. In Harvard Business Review’s 2011 study of 1.25 million leads, firms that tried to contact leads within an hour were nearly seven times as likely to have a meaningful conversation with a decision maker as firms that tried an hour later, and more than 60 times as likely as firms that waited 24 hours or longer.

How do I calculate contact rate in a spreadsheet?

Keep one row per unique lead with a created date and a first right-party contact date. Add a helper column that returns 1 if the contact date exists and is within 14 days of the created date, otherwise 0. Contact rate is the sum of that column divided by the count of leads for the cohort, calculated only once the youngest lead is 14 days old.