When Local Businesses Actually Answer — And Why the Published Benchmarks Contradict Each Other

Search for cold call connect rates and you will find, published this year, with equal confidence: 2–3%, 4–6%, 8–12%, and 15–28%. Not a range within one study — four different central claims from four sources, each presented as the benchmark.

The instinct is to average them, or to pick the flattering one. Both are mistakes. The numbers differ by an order of magnitude because they are measuring different denominators: dials versus contacts, raw lists versus verified direct dials, SDR teams calling enterprise switchboards versus someone ringing a plumber's mobile. A "connect rate" without its denominator is not a statistic. It is a number with a percentage sign.

This matters more than it sounds, because teams set targets against these figures. A local-market caller benchmarking against enterprise SDR data will conclude they are failing when they are not — or, worse, conclude they are succeeding when their list is quietly the reason.

What the Denominators Actually Are

Four measurements hide behind one phrase, and the gap between them explains nearly all the published disagreement.

Dial-to-meeting is the harshest and the most quoted out of context. Roughly 100–150 dials to book one B2B meeting is a common figure. It sounds catastrophic until you notice it compounds every other rate in the funnel, including the quality of a list you may have built badly.

Dial-to-conversation sits around 8 dials per conversation. This is the number that measures reachability, and it is the one a local-market caller should actually track, because it isolates "can I get this person on the phone" from "is my pitch any good."

Conversation-to-meeting runs near 12 conversations per meeting. This is the number that measures your opener, your offer and your targeting. If this is bad, calling at a different hour will not save you.

Attempts-to-reach is typically 6–10 attempts before a given prospect picks up at all. This is the one most teams violate: nearly half of reps stop after a single attempt, which means they are sampling the small fraction of prospects who answer immediately and drawing conclusions about the rest.

Read that way, the contradictory headline numbers resolve. A team reporting 15–28% is measuring dial-to-conversation on verified direct dials. A team reporting 2–3% is measuring dial-to-meeting on a scraped list. Neither is lying.

Timing: Real, Overstated, and Different for Trades

The timing research is genuinely useful and routinely oversold. The consistent finding across 2026 datasets is that mornings win — roughly 8–11am local time, with Tuesday through Thursday outperforming Monday and Friday, and Tuesday and Wednesday together accounting for something like 44% of demos booked. A second, weaker window appears in the late afternoon.

Three caveats keep this from becoming superstition.

The effect is a lift, not a gate. Calling in the right window is reported to lift connect rates by something like 30–50% relative to the wrong one. That is worth having. It is not the difference between a working channel and a broken one, and no amount of timing optimisation rescues a list of businesses that have no reason to talk to you.

"Local time" is doing a lot of work. The single most common self-inflicted wound in multi-region calling is running a whole list on the caller's clock. If your prospects span time zones and your dialler does not respect theirs, your carefully chosen 10am window is 7am for a third of the list.

Local trades do not keep office hours, and the generic advice partially inverts. The 8–11am rule comes from datasets dominated by desk-based B2B. A roofer is on a roof at 10am. A restaurant is in service at noon and dead at 3pm. A dental practice has a receptionist all day and a decision-maker available almost never. A hairdresser's quiet hour is Tuesday morning; their impossible hour is Saturday. The published window is a prior, not an answer — and the answer differs per trade in ways that are learnable from your own call log within about two weeks.

Where Reachability Is Actually Won

The uncomfortable finding in the data is that timing is a second-order lever. The first-order levers are unglamorous.

Number quality dominates everything. Verified direct-dial numbers are reported to lift connection rates by up to 40% — larger than any timing effect. For local businesses this usually means the number on the website contact page rather than the one on the aggregator listing, and it means checking that the number still belongs to the business at all. A scraped list where 15% of numbers are stale does not have a timing problem.

Attempt discipline beats attempt volume. If it takes 6–10 attempts to reach someone, a list worked once is a list you have not tested. The teams with poor reachability are frequently not calling badly; they are calling each prospect 1.4 times and calling that a campaign.

Gatekeepers are a routing problem, not an adversary. In local business the person answering is often the owner's spouse, the practice manager, or whoever is nearest the phone. They are not screening you out of policy; they are triaging. A specific, checkable reason for the call routes past them because it gives them something to relay. "I'm calling about your website booking page" is relayable. "I wanted to speak to the owner about growth" is not.

Reachability decays and nobody notices. Business phone numbers rot at a few percent a month. A list sourced in January and worked in June has a materially different connect rate for reasons that have nothing to do with your calling. Without a capture date on every record, you will attribute that decay to your script.

What to Do With Your Own Numbers

The externally published benchmarks are useful for exactly one thing: telling you which quantity to measure. Beyond that, your own log outranks them, because it is the only data measured on your list, your market and your denominator.

Log the denominator, not just the outcome. Record dials, conversations and meetings as three separate counts. A single "connect rate" field is how teams end up unable to tell a list problem from a pitch problem — the two have opposite remedies and identical symptoms in an aggregate number.

Segment call outcomes by trade before you segment by hour. With a few hundred calls you will find that dentists and roofers have different answerable windows, and the difference between trades will be larger than the difference between morning and afternoon within a trade. Optimising the hour before you have segmented the trade is optimising noise.

Set the attempt policy in the system, not in the rep's judgement. Decide the number of attempts and the spacing, then make the pipeline surface the next attempt automatically. Attempt discipline is the single highest-return change available to most local outbound teams, and it is the one most reliably lost to human optimism about which prospects are "not interested."

Treat a stale number as a data defect, not a bad call. When a number is disconnected, that outcome should update the record, not just the call log. Otherwise the same dead number is redialled by the next person to work the list, and your connect rate quietly encodes your data hygiene.

Re-baseline after any list change. Comparing this month's connect rate to last month's is only meaningful if the list came from the same source with the same verification. Most apparent performance swings in small outbound teams are list-composition changes wearing a performance costume.

Cold calling local businesses works, and it works at rates that would look implausible to someone whose mental model comes from enterprise SDR benchmarks. But the published numbers cannot tell you that, because they are not measuring your funnel. The useful move is not to find the correct benchmark. It is to instrument your own three denominators, work each prospect the six-to-ten times the data says it takes, and let two weeks of your own call log tell you when a roofer picks up the phone.