Measuring Your Real Internet-Lead Response Time (A How-To)

Speed to contact is one of the few variables in the internet sales process that a dealer controls completely and can change overnight. Yet most response-time discussions stay at the level of folklore rather than measurement. This guide skips the motivation speech and goes straight to the mechanics.

Why does your CRM report a different number than reality?

CRM systems typically record response time as the gap between lead creation and the first logged activity. That activity is often a status change, an internal note, or an auto-responder trigger, none of which represent contact with the customer. The number looks good in a Monday morning report and tells you almost nothing about the buyer's experience.

The gap between logged activity and actual human contact typically comes from three sources: leads routed to a shared inbox no one monitors continuously, CRM tasks created but not worked immediately, and after-hours leads that sit until the next business morning without any escalation rule in place.

How do you pull the raw timestamps you actually need?

Accurate measurement requires three data pulls run against the same lead population, then joined on a common identifier such as the lead's email address or a provider-assigned lead ID.

Step 1: Lead provider delivery timestamp. Every major lead provider, whether it is your OEM portal, a third-party aggregator, or a website form processor, logs the exact UTC time the lead was submitted and the time it was delivered to your endpoint. Request a 30-day export in CSV format. The column you want is typically labeled "lead received" or "submission timestamp," not "delivered to dealer," because delivery latency itself can be a hidden problem.

Step 2: CRM first human-activity timestamp. Export all lead records for the same 30-day window. Filter activity types to exclude: auto-responder sent, lead created, lead assigned, and status changed. Keep only: outbound email manually sent, outbound call logged, and outbound text sent. The earliest qualifying timestamp per lead is your CRM response time. If your CRM does not let you filter by activity type, your measurement problem is a configuration problem first.

Step 3: Email server logs. Your email platform, whether it is Google Workspace, Microsoft 365, or a dealership-specific mail host, maintains send logs with per-message timestamps. Export sent-mail logs for your BDC or internet team addresses for the same period. Match on lead email address. This cross-check catches cases where a salesperson replied from a personal email account that never synced back to the CRM.

Once you have all three files joined, calculate two numbers: median elapsed time and 90th-percentile elapsed time. The median tells you your typical performance. The 90th percentile tells you how bad things get for your worst-served customers, which is often the more operationally important number.

What does the research say about how fast is fast enough?

The most-cited study on this topic remains the Harvard Business Review audit conducted by James Oldroyd and colleagues, published in 2011, which found that contacting a lead within one hour made a company nearly seven times more likely to have a meaningful conversation than contacting the lead even one hour later. Automotive-specific follow-up research from Marchex and from the Haig Report has consistently shown median dealer response times measured in hours, not minutes, despite the one-hour threshold being industry knowledge for over a decade.

For context: a five-minute median response during business hours is achievable with dedicated staffing and a clean routing setup. After-hours leads require a different strategy entirely, whether that is a monitored chat function, a live transfer service, or a structured next-morning callback queue with escalation rules that fire at 8 a.m. sharp.

How do you segment the data so it tells you something actionable?

Aggregate response time numbers hide the problems. Segment your results by the following dimensions before drawing any conclusions.

• Hour of submission: Leads submitted between 9 a.m. and 5 p.m. versus leads submitted outside business hours. These two populations have different response-time distributions and require different fixes.
• Lead source: OEM leads, third-party leads, and your own website leads may route differently and show different lag patterns.
• Day of week: Saturday afternoon leads at many stores sit until Monday. That is a policy problem, not a staffing problem.
• Salesperson or BDC agent assigned: Individual performance distribution is almost always wider than managers expect. Your fastest responder and your slowest responder are often in the same building.
• Vehicle type or price band: Some stores triage leads informally by perceived deal size. The data will show this if it is happening.

Once segmented, rank the combinations by 90th-percentile response time, longest first. The top three to five combinations on that ranked list are your intervention priorities. Everything else is maintenance.

What are the most common root causes and their fixes?

After the measurement is clean, root causes tend to fall into a short list.

Shared inbox without an owner. Leads arrive in a group email address that everyone assumes someone else is watching. Fix: assign a named, rotating monitor with a defined handoff protocol and a maximum unattended window of 15 minutes during business hours.

CRM notification routing to a role, not a person. Some CRM configurations send new-lead alerts to a role-based user rather than a live individual. Fix: audit your routing rules and confirm that every lead in every hour block has a named human on call.

After-hours leads with no escalation rule. Leads submitted after close sit until open with no automated escalation. Fix: build a workflow that flags any lead older than 30 minutes submitted between 6 p.m. and 8 a.m. and queues it for first-call priority when the store opens, with a manager notification if it has not been contacted within 20 minutes of open.

Manual note-taking without a call log. Salespeople call but do not log, so CRM response time shows hours when actual contact happened in minutes. Fix: enforce click-to-call or post-call logging as a condition of CRM access. Unlogged calls are invisible calls.

For a broader look at where internet lead programs break down, see our AI reality check and the lead intelligence overview.

If you want a structured outside review of your current setup before building new workflows, our diagnostic call is a 45-minute working session, not a sales pitch. More on the full scope of what we audit is on the services page.

The MIT Sloan Management Review has also published useful follow-on work on lead response decay curves if you want additional academic framing: see their research index at mitsloan.mit.edu.

What should you do with the measurement once you have it?

Measurement is only useful if it changes behavior. Build a simple weekly scorecard: median response time by agent, median response time by lead source, and percentage of leads contacted within 30 minutes. Post it. Review it in the Monday meeting. That is the minimum viable accountability loop.

Do not set a target before you know your baseline. Setting a five-minute target when your current median is 4.5 hours creates the wrong kind of pressure and typically produces data manipulation rather than process improvement. Set the first target at a 50 percent reduction from your current median, hold it for 30 days, then step down again.