How ReplyClock tests lead response

ReplyClock measures how each location of a multi-location brand handles a new inquiry sent by web form, email or site chat. We send a realistic service question, clearly labeled as a ReplyClock test, time the reply, and grade it with one published rubric.

Disclosed and permissioned, never covert

This is disclosed web-lead testing, not covert mystery shopping.

• Every inquiry, including a pre-client benchmark, is sent only with written permission.
• Client tests follow a scope you approve in writing.
• Each inquiry says it's a ReplyClock response-time test and asks one realistic question.
• We never book appointments. We never reply after your team's first message, so we don't steer the follow-up we're measuring.
• We never use solving tools or workarounds for a CAPTCHA. A visible challenge is completed only by hand by a ReplyClock team member; if no one can, the test is marked “not testable” and left out of every average.

What each test is scored on (100 points)

• Speed to first useful reply (25): business-hours time from submission to the first reply, from a person or an AI, that engages the question. A generic autoresponder doesn't count.
• Was the question answered? (25): did the reply answer it or deflect it?
• Booking path offered (15): a booking link or a specific time scores best. "Let us know" scores least.
• Follow-up over 7 days (15): follow-ups that add something or move toward a booking count fully. Repeated identical nudges count as an automated drip.
• Personal and specific (10): does the reply speak to what we asked and where we asked it?
• Professionalism (10): the right name, brand and location; no errors; a named sender with contact details; a courteous tone; no made-up details.

Testing AI reply tools

We score AI and human replies on the same rubric. Each reply is labeled likely person, likely AI or automated, or unclear, with our reason. That label is our judgment and never changes the score. We also note whether an automated reply handed the lead to a named person, and we deduct for details a bot made up. Qualifying questions are noted, not penalized.

How results are reported

• One outcome per test: useful reply, autoresponder only, confirmed delivered with no reply, submission failed, or not testable.
• Intake health, separately: a form that errors out, or a channel we can't test honestly, is reported on its own line. It's never averaged into the response score.
• Confidence on every score: each score shows its sample size (n). One test is a "single sample". 2–3 tests are an "early read". 4 or more across at least 2 days and 2 channels are a "pattern". Only pattern locations are ranked.
• Trends: we describe a location as "improving" or "declining" only when its score moved 10 points or more and both months have at least 2 tests. Otherwise we report the scores without a trend label.
• Fixes and retests: scorecards list corrective actions. When a fix is marked done, a retest can be scheduled within your plan to record whether it held.
• Evidence: we keep timestamps and the evidence that's relevant to each channel, such as the reply and its details, subject to privacy limits.

When a form can't be tested

Some forms don't offer a suitable way to ask a question, for example a form that only books appointments. Those may be marked "not testable" rather than scored. Timing and channels are agreed in writing before testing.

What we don't measure

Phone calls and text messages. We also don't work with healthcare or medical aesthetics, facility services, schools, or cleaning and janitorial brands.

Plans

Starter $150/month, Growth $400/month, Network $800/month. Every plan uses the same rubric and evidence standard. See pricing for what each plan includes.

See it for yourself

Want to see how your locations do? With your written permission, we'll test three of them for free and send you a scorecard with grades, timings and fixes. You can also view the sample scorecard, which uses example data for a fictional brand. Email hello@replyclock.com or call (817) 668-6981.

Send us your brand name and three locations, and we'll set up your free snapshot.