What Is a Lead? Why Conversations, Not Forms, Define Customer Intent

Customers do not buy vehicles by filling out forms. They buy vehicles through conversations. A lead is not defined by how it enters the dealership. It is defined by whether the customer engages.

---

Executive Summary

For decades, the automotive industry has measured success by counting leads. Website leads, third-party leads, phone leads, chat leads. Every CRM, dashboard, and marketing report begins from the same assumption: a lead is a lead. This report tests that assumption against 1,042,103 automotive leads handled by Strolid’s business development teams for its franchised dealership clients during the 2024 and 2025 calendar years, and against the roughly 7.5 million calls, emails, texts, and chats exchanged with those leads.

The assumption fails. Whether a lead becomes a customer tracks less with its source, its channel, or the volume of follow-up it receives than with whether the customer ever participates in a genuine two-way conversation.

In the data, 64.1% of leads engaged in at least one two-way conversation, a live call with a substantive outcome, a text or email reply in the customer’s own words, or a chat. Measuring engagement before any appointment existed, leads with such a conversation went on to book at 28.2%, show at 16.4%, and purchase at 8.1%. Leads with no such conversation booked at 11.0%, showed at 7.1%, and purchased at 3.5%. That is an advantage of two and a half times on appointments that holds at better than twice the rate at every stage of the funnel, and it is not an artifact of a few stores: at 98% of the dealerships with enough volume to measure, engaged leads out-booked silent ones, with a median advantage of 3.0 times per store. Even under the strictest possible tests, which remove every appointment-setting conversation from the definition of engagement or reclassify it on the most conservative clock, the engaged group still converts at 1.3 to 1.6 times the rate. Measured over the full life of each lead, engaged leads accounted for 89.5% of all appointments and 85.5% of all sales.

Engagement also has depth. Grading every lead by how far its pre-appointment conversations progressed, from no participation to an appointment-stage exchange, the sale rate climbs at every level: 3.5%, 4.5%, 8.2%, 20.9%. These are associations along the path a buyer travels, not isolated causal effects, but they mean the depth of a lead’s conversation is a running read on where the customer stands, one that no lead form carries.

Just as telling is what does not track conversion. Among roughly one million leads that received follow-up email, the leads that opened emails but never replied purchased at 4.0%, no better than leads with no recorded open at all (4.3%). The leads that replied purchased at 10.6%. Opens are activity; replies are engagement. In this data, attention without participation carried no lift, and only participation moved with the outcome.

Engagement is also perishable. Half of all engaged leads had their first two-way conversation within about an hour of the lead arriving, and 81.6% within the first day. Leads whose first conversation happened in that first hour went on to purchase at 10.2%; leads that engaged after a week purchased at 4.6%.

The conclusion is a redefinition. Counting leads measures how many names entered the CRM. Counting conversations measures how many customers are actually in motion. Dealerships that measure engagement, the rate, the speed, the depth, and the signals inside each conversation, will see their pipeline as it actually is, prioritize the customers who are actually participating, and put themselves in a far better position to sell more vehicles. The conversation is the asset. Everything else is the channel it happens through.

---

1. Introduction: The Lead-Counting Habit

Ask a dealership how last month went and the first number offered is almost always lead volume. Ask how the marketing budget is allocated and the answer is cost per lead. The lead count is the industry’s founding metric, and everything downstream, CRM dashboards, vendor scorecards, provider reports, inherits it.

The habit made sense when a lead was scarce and expensive. But the modern lead is neither. Today a “lead” can be a hand-raised buyer who wants an appointment this Saturday, or a mistyped email address attached to an abandoned browsing session. Both enter the CRM. Both are counted. From a reporting perspective, every lead begins life equal. From the customer’s perspective, they are anything but.

The industry has responded to rising lead volume mostly by industrializing the response. That work is real and measurable: in DAS Technology’s lead-response study of more than 1,700 dealerships, presented at NADA 2025, 61% of dealerships responded to internet leads within 15 minutes, up from 55% two years earlier.^[1] Pied Piper’s 2026 Internet Lead Effectiveness study recorded the industry’s best-ever responsiveness score, with 62% of dealerships following up by both phone and text.^[2] Yet the same research shows the floor remains low, with roughly one dealership in five failing to respond personally to its own website customers at all.^[2:1]

Response, however, is only half of a conversation. A response measures what the dealership sent. It says nothing about whether the customer ever answered. And the customer’s answer, the reply, the question, the callback, is where buying actually begins. Cox Automotive’s Car Buyer Journey research finds that buyers overwhelmingly prefer a mix of digital and human interaction, with 63% calling an omnichannel process ideal;^[3] Urban Science and The Harris Poll find that 82% of consumers consider dealership follow-up important to their purchase decision.^[4] The Marchex Institute, analyzing more than 307,000 dealership calls, found that 57% of inbound sales calls carry purchase intent.^[5] Customers signal intent by conversing. The question is whether dealership measurement captures it.

This report proposes, and then tests, a different founding metric. Rather than asking how many leads arrived, it asks how many customers engaged, and what that engagement was worth. It shows that the conversation itself, its existence, its depth, its speed, and its content, is the most direct measure of customer intent available in dealership data.

---

2. What Engagement Is, and Is Not

Before engagement can be measured, it has to be defined, and the definition has to exclude the things dealership reporting has historically confused with it.

Customer engagement is a meaningful two-way conversation in which the customer actively participates and demonstrates where they stand in the buying process.

The key word is participates. A customer is not engaged because a dealership attempted contact. A voicemail left, an email delivered, a text sent, even an email opened: these are activities of the dealership, not actions of the customer. They prove effort. They do not prove interest.

An engaged customer does something only a customer can do: answers the phone and talks, replies to a text or email in their own words, asks a question about a vehicle, discusses payments or a trade, requests a callback, starts a chat, schedules or discusses an appointment. Each of those acts is voluntary participation in the buying process, and each adds information about where the customer actually is in it. Usually what participation reveals is continued interest; sometimes it reveals the opposite, a customer who bought elsewhere or left the market. Both are engagement, and both are worth knowing, which is why the measurement counts any meaningful two-way conversation, whatever its direction.

Engagement also has depth. A customer who says “please call me after work” has engaged, but a customer negotiating payments on a specific vehicle has engaged at an entirely different level. This report grades every lead on a four-level scale, applied to the conversations that occur before the lead’s first appointment:

• Level 0, No recorded engagement. No two-way conversation is recorded before the lead’s first appointment, or ever, for leads that never reach one.
• Level 1, Initial engagement. The customer responds once, briefly: an acknowledgment, a request to be called later, a short exchange that does not yet go anywhere.
• Level 2, Active engagement. The customer participates in a sustained way: multiple two-way interactions, or at least one substantive conversation long enough to discuss vehicles, payments, trade, or timing.
• Level 3, Appointment-stage engagement. The conversation itself reaches a purchase step: an appointment is set or confirmed live in the exchange.

Engagement answers the question: did the customer actively participate in a meaningful conversation? Engagement depth answers a different question: how far into the buying process has that participation gone? Section 4 measures both against 1.04 million real outcomes.

---

3. Methodology

Data source. The analysis was computed directly from Strolid’s source operational data warehouse: the raw lead, appointment, and sale records generated by Strolid’s BDC work for franchised dealerships, joined to Strolid’s conversation records, call records with dispositions and durations captured as vCons (the IETF-standardized conversation container Strolid created), customer text and email replies, and chat transcripts. It does not rely on a pre-summarized reporting layer.

Cohort. All leads created in the 2024 and 2025 calendar years (January 1, 2024 through December 31, 2025) across the dealerships subscribed to Strolid’s full BDC service, with administrative, junk, and unworked lead records removed (bad, spam, duplicate-merge, and inactive codes among them). The cohort contains 1,042,103 leads. Funnel outcomes are computed on a distinct-lead basis: an appointment means the lead has at least one active, non-deleted appointment record, a show means the lead has an appointment marked as shown, and a sale means the lead is marked sold. The cohort produced 226,341 appointments (21.7%), 134,443 shows (12.9%), and 66,582 sales (6.4%).

What counts as an engaged conversation. A lead is engaged when at least one of the following exists in its record: (a) a call whose logged outcome implies a two-way exchange, an appointment set, declined, or confirmed, a callback scheduled, a transfer, an escalation, a stated “not in market” or “bought elsewhere,” or similar dispositions a specialist can only record after talking to the customer; (b) an answered call lasting at least 60 seconds; (c) a text or email reply from the customer captured in the CRM reply record; or (d) a chat conversation. Voicemails, unanswered and missed calls, bad numbers, hang-ups, outbound sends, and email opens do not count. The cohort contains 2,513,204 engaged customer interactions: 1,176,901 qualifying calls, 1,318,058 customer text and email replies, and 18,245 chats. (Reply and chat records are counted as individual interaction events, not sessionized dialogues, which is why this report speaks of interactions rather than a count of distinct conversations.)

Measuring before the outcome. The headline comparison in Section 4.2 and the levels in Section 4.3 are anchored on the lead’s first appointment record: they count only conversations that occurred before it existed (for leads that never reached an appointment, all their conversations count). Conversations that follow an appointment, confirmations, reschedules, feedback calls, are excluded from the predictor, because an appointment generates conversations of its own. Show and sale rates are then read downstream of that same anchor; a sale-date-anchored cutoff is possible only at day granularity and is reported as a bracket in Section 4.2. Two measurement choices remain, and both are disclosed with sensitivities. First, a conversation in which the appointment is set is still a conversation, and it is counted; removing even those leaves the advantage at 1.4 to 1.6 times. Second, call timestamps and appointment timestamps are recorded on clocks roughly five hours apart, so classification near the boundary is uncertain for about 5% of engaged leads; re-running the comparison on the most conservative clock alignment, which also pushes the appointment-setting call itself past the boundary, leaves 1.3 to 1.5 times. The true advantage lies between those floors and the primary figures.

Engagement levels. Assigned from pre-appointment conversations as defined above. Level 1 is exactly one brief engaged interaction; Level 2 is two or more interactions or at least one answered call of two minutes or longer; Level 3 is a conversation whose recorded outcome is an appointment being set or confirmed.

Scope and limitations. This is an observational analysis of operational data, not a controlled experiment; engaged customers are partly self-selected, and Section 5 addresses directly what the data can and cannot say about causation. Conversation records must be linked to a lead to be counted, and linkage is imperfect, especially for 2024 calls, so some genuinely engaged leads are recorded here as unengaged; if that misclassification is unrelated to outcomes it attenuates the measured gaps, but its direction and magnitude were not quantified. Email-open tracking does not cover every email sent. The sale date is recorded at day granularity, so sale-anchored cutoffs are reported as a bracketed range (Section 4.2), and the appointment record, which carries a precise timestamp, anchors the headline comparisons. Call timestamps and lead timestamps are recorded on different clocks several hours apart; timing analyses use a consistent timestamp source. The AI conversation tags used in Section 4.7 exist only from May 2025 onward and are analyzed as a disclosed subset. Some of the dealerships in the cohort run service-heavy operations; their leads rarely enter the sales funnel. Because the underlying tables are live, counts may drift modestly on re-run; figures reflect the data as analyzed in July 2026.

---

4. Findings

4.1 The conversation landmarks

Between the lead and the sale sits a sequence most dealership reporting never measures. Adding engagement to the standard funnel makes the middle visible. The table reports each milestone as a share of all leads; the stages overlap rather than nest perfectly (some appointments, for example, come from leads with no recorded conversation), so it should be read as a set of cohort landmarks rather than a strictly sequential funnel.

Nearly two leads in three become a conversation. Just under half the cohort reaches sustained, active engagement. More than one in five books, one in eight shows, one in sixteen buys. The largest single drop is at the very first step: more than a third never converse at all. The industry measures the top line and the bottom three lines of this table. The two lines it skips are where the outcome takes shape.

Producing those interactions took work at scale: roughly 7.5 million communication events across the cohort, 3.0 million calls handled, 2.8 million emails, 1.7 million texts, and 18 thousand chats, which yielded 2.5 million engaged customer interactions. Engagement is also not something the dealership does alone. In 58.1% of engaged leads, at least one interaction came from the customer’s side: 18% placed an inbound call themselves, and 48% wrote back in their own words by text, email, or chat (the two groups overlap). The customer, given a reason, leans in.

4.2 Engagement changes the odds

The central comparison in this report asks whether a lead engaged in a two-way conversation before any appointment existed, and then follows both groups through the funnel.

A lead that engages in a real conversation converts at well over twice the rate of a lead that never does, and the advantage holds all the way through the funnel.

Because a conversation that sets an appointment counts as engagement, part of the appointment-rate gap is definitionally linked, and because call clocks and appointment clocks differ by several hours, classification near the boundary is uncertain for about 5% of engaged leads. Two sensitivity checks bound both concerns. Stripping every appointment-setting conversation out of the engaged definition entirely leaves the engaged group ahead at 1.4 times on appointments, 1.5 times on shows, and 1.6 times on sales. Re-running the split on the most conservative clock alignment, which also reclassifies the appointment-setting call itself as post-appointment, leaves 1.3, 1.4, and 1.5 times. The advantage is not an artifact of labeling or clocks; the primary figures and these floors bracket it. Anchoring on the sale date instead of the appointment is possible only at day granularity and brackets the sale advantage between 1.1 and 2.8 times.

Measured over the full life of each lead, engaged leads accounted for 202,683 of the cohort’s 226,341 appointments (89.5%) and 56,942 of its 66,582 sales (85.5%). The remainder may largely reflect conversations the data cannot link to the lead and appointments set outside the BDC’s recorded channels, walk-ins and dealer-set appointments among them.

The pattern is also not carried by a handful of large stores. Among the dealerships with enough volume in both segments to measure, pre-appointment engaged leads out-booked silent leads at 98% of them, with a median per-store advantage of 3.0 times; even on the strict-clock variant, engaged leads still out-book at four stores in five, with a median of 1.4 times.

4.3 The engagement gradient

Grading each lead on the four-level scale from Section 2 turns the binary split into a dial. The levels describe how far a lead’s pre-appointment conversations progressed; the rates that follow are associations along that progression, not isolated causal effects of each conversation.

The sale rate climbs at every step: 3.5%, 4.5%, 8.2%, 20.9%. (Appointment rates are omitted from the Level 3 row by design: Level 3 is defined by a conversation that sets an appointment, so its 96% appointment rate is definitional rather than evidential. Its show and sale rates are downstream of that definition, and they are the highest in the table by a wide margin: nearly half of Level 3 leads physically showed, and one in five bought.)

Two features deserve emphasis. First, Level 1 is worth surprisingly little. A customer who responds once, briefly, and is not brought into a real conversation goes on to buy at 4.5%, about seven-tenths of the cohort average, despite having actually responded. Getting a response is not the finish line; it is the starting line. Second, the big step is Level 1 to Level 2, where the sale rate nearly doubles. In this data, the largest single difference in the odds comes not with contact but with the first sustained, substantive conversation.

4.4 Activity is not engagement

If engagement were just a proxy for “the dealership worked the lead harder,” then measures of activity should track conversion the way conversation does. Email offers a direct test, with two disclosed limits: open tracking does not cover every email, so “no recorded open” includes some opens the system never saw, and the reply group includes customers who replied by text as well as email, measured over the lead’s lifetime.

Among the 982,862 cohort leads that received at least one BDC follow-up email:

The leads that opened emails but never answered converted no better than the leads with no recorded open, on these figures, slightly worse. The leads that replied purchased at roughly two and a half times the rate of either, and booked at roughly twice the rate. Within the limits of the tracking data, the pattern is hard to miss: an email open is a fingertip on a screen, a reply is a customer talking, and only the reply moved with the outcome. Dashboards that celebrate open rates are measuring the dealership’s reach, not the customer’s interest.

The same logic explains a finding inside the conversation data itself. Among leads that engaged before their first appointment, the leads with exactly one conversation actually booked at the highest rate, 38.0%, because when the conversation is substantive, the appointment is very often set right there, in the first real exchange, and no further pre-appointment conversation is needed. Depth of participation, not volume of touches, is what tracks outcomes in this data: one substantive conversation carries more signal than any number of deliveries, and often more than a second conversation.

4.5 Engagement compounds across channels

Engagement is not a single-channel behavior. Among engaged leads, the mix of channels the customer used over the lead’s life tracks the outcome:

A customer who both talks on the phone and writes back buys at double the rate of a phone-only customer and roughly 1.5 times the rate of a messages-only customer. Much of this is self-selection, more committed buyers naturally use more channels, but the operational implication holds either way: a customer active in two channels is among the strongest signals in the dataset, and multi-channel conversations are worth inviting and routing on, not just permitting. This echoes Pied Piper’s finding that the industry’s best lead handlers now follow up by phone and text together.^[2:2]

4.6 Engagement is perishable

Engagement, when it comes, comes fast. Half of all engaged leads had their first two-way conversation within about 50 minutes of the lead’s arrival, and 81.6% within the first day. And the earlier it comes, the more it is worth:

A further 480 leads (0.07% of engaged) with timestamp anomalies are excluded from this table.

A lead whose first real conversation happens in the first hour purchases at more than twice the rate of one whose first conversation happens after a week. The direction matches the classic speed-to-lead research, Harvard Business Review’s audit found firms responding within an hour were nearly seven times likelier to have a qualifying conversation than those waiting even an hour more^[6], but the metric here is stricter than response time. What decays is not the window to send something; an automated response can always be instant. It is the window in which the customer will still talk.

4.7 Conversations reveal intent, in the customer’s own words

Everything above measures the existence, depth, speed, and channel of engagement. The content of the conversations is a further layer, and since May 2025 Strolid’s AI conversation intelligence has tagged that content at scale. For cohort leads inside that window, the tags read like a map of buying intent that no lead form could carry:

These signals track outcomes. Among engaged leads created in the second half of 2025, leads whose conversations carried at least one buying signal, a question, a callback request, an ordering interest, a finance or trade discussion, booked appointments at 42.8% against 27.3% for engaged leads without one, and purchased at 11.5% against 9.0% (appointment-setting tags excluded from this comparison to avoid circularity; tags are matched over the lead’s lifetime, so the comparison is descriptive rather than strictly predictive). The conversation does not just indicate that a customer is engaged. It tells the dealership what the customer needs next, an answer, a callback, a payment figure, an appraisal, and just as valuably, it surfaces the risks: the competitor mentioned, the price objection, the purchase delayed. Historically, all of this lived and died inside individual phone calls. Captured and structured, it becomes operational intelligence: which conversations need a manager, which need a trade appraisal, and which need immediate action because the customer is ready now.

---

5. Analysis: What Engagement Proves, and What It Doesn’t

5.1 Signal, selection, and what the data can say

The honest objection to everything above is selection: engaged customers were more interested to begin with, so of course they buy more. That is partly true, and it is worth saying plainly. Engagement is not a lever that mechanically converts any lead two-and-a-half-fold, and this observational data cannot fully separate how much of the gap is the conversation doing work and how much is interested customers revealing themselves. The distinction matters for expectations: if engagement is mostly a signal, then inducing one extra conversation from a low-intent customer will not automatically reproduce the gaps reported here.

What the data does establish is narrower but still consequential. First, participation, not attention, is where the information lives: opened emails carried no measurable lift while replies carried a large one (Section 4.4), so the signal is specifically the customer’s act of participating, not their exposure to follow-up. Second, mere contact is not the signal either: a single brief response (Level 1) barely moves the outcome, while sustained conversation (Level 2) nearly doubles it. Whatever engagement measures, it is measured in depth, not in touches. Third, the gradient persists within phone-presence strata. Within leads that arrived with a phone number, pre-appointment engagement still carries a 1.9-times sale advantage; within the minority that arrived without one, the small group that nonetheless engaged out-sold its silent peers by roughly five times. Stratifying by phone presence controls contactability, not underlying intent, but it rules out the simplest confound, and the sensitivity checks in Section 4.2 show the advantage is not manufactured by labeling or clocks.

Operationally, both readings point the same way. If conversation converts, dealerships should create more conversations. If conversation reveals, dealerships should create more conversations, because the revelation is the most valuable prioritization data they can own, and it only exists once the conversation happens. What the two readings price differently is the expected return, which is why the recommendations in Section 6 treat engagement as the thing to measure and act on, not as a guaranteed multiplier.

5.2 Against the industry backdrop

The industry’s own research points the same direction from three angles. Pied Piper and DAS Technology document an industry that has industrialized responding, faster responses every year, more channels per response, while roughly one store in five still fails to respond personally at all.^[2:3][1:1] Marchex found 57% of inbound sales calls carry purchase intent, and that sales calls convert to purchases at roughly four times the rate of email, evidence that when the customer chooses a live channel, intent is concentrated there.^[5:1] And consumer research from Cox Automotive and Urban Science shows buyers want the human exchange at the decisive moments.^[3:1][4:1] What this report adds is the measurement layer those studies imply but do not build: a definition of engagement grounded in customer participation, and a demonstration across 1.04 million leads that participation carries the signal that activity metrics, where this data could test them, did not.

5.3 The measurement error hiding in plain sight

If a lead is a lead, then a dealership evaluating lead sources, ad spend, or BDC performance by volume and cost-per-lead is optimizing an input that explains little. Two sources delivering 1,000 leads each are not equal if one delivers 700 conversations and the other 300: at the engagement-segment sale rates measured here, that difference is associated with roughly 40 percent more sales from the same lead count. The lead count is not wrong; it is just the wrong denominator for many of the decisions that matter. Cost per engaged conversation, appointments per engaged conversation, and sales per engaged conversation are the versions of those metrics that align with how customers actually buy.

---

6. From Lead Management to Conversation Management

The findings translate into six concrete moves for dealers, BDC leaders, and dealer groups.

1. Make engagement rate a headline KPI. Track the share of leads that reach a genuine two-way conversation, 64.1% in this data, alongside lead volume. It is the leading indicator that sits closest to appointments and sales in this analysis, and most reporting stacks do not compute it at all.

2. Measure speed-to-conversation, not just speed-to-response. Auto-replies answer in seconds; what pays is the first real exchange. Half of engagement happens inside the first hour, and first-hour conversations convert at twice the rate of first-week ones. Staff, route, and escalate accordingly.

3. Treat the first substantive conversation as the event that wins the appointment. The data’s biggest single lift is from brief contact to sustained conversation (Level 1 to Level 2, sale rate nearly doubling). Train specialists to convert an acknowledgment into a discussion, of the vehicle, the payment, the trade, in the same exchange, because often there is no second one: when the first conversation is real, the appointment is usually set right there.

4. Invite multi-channel participation. Customers who both talk and write back convert at the highest rates in the data. Every call should open a text thread; every text thread should offer a call. Some of that gap is self-selection, but the routing implication is unaffected: a customer answering in two channels is a priority customer.

5. Stop crediting activity metrics with intent. Opens, deliveries, and dials are effort measures, useful for managing workload; email opens, the one activity signal this data could test directly, carried no conversion lift at all. Reserve intent language for participation: replies, answered conversations, questions, callbacks.

6. Mine conversations for signals, and route on them. A finance question, a trade mention, a callback request each mark a customer telling the dealership what the next step is; a competitor mention or price objection marks one about to leave. Whether captured by AI or by disciplined disposition notes, these signals should drive prioritization and manager involvement the same day they occur, because the conversation that revealed them is already aging.

---

7. Looking Forward: More Conversations, Not Fewer People

Artificial intelligence is changing how dealerships communicate, and the temptation will be to point it at the metric the industry already worships: response volume. That would automate the wrong thing. The evidence in this report says the scarce, valuable event is the engaged conversation, and the highest use of AI is to create more of them, and to understand all of them.

To be explicit about scope: this dataset measures human conversations, and it does not test AI-driven engagement channels. What it does establish is the standard any new channel must meet, genuine customer participation. An AI copilot that answers a customer’s question at 11 p.m., provides vehicle information, estimates a trade, or books the appointment, and then hands the customer to a person the moment judgment or commitment is involved, would add a real engagement channel by that standard. Conversational intelligence that reads every call, text, and chat, surfaces the buying signals, and flags the at-risk conversations turns the dealership’s largest untapped dataset into operating decisions; Section 4.7 is an early measure of exactly that. What AI should not do is replace the conversation it was meant to create: a decade of consumer research shows buyers still want the human exchange at the decisive moments of a purchase, and every gradient in this report is consistent with that.^{[3:2][4:2][7]} Strolid’s operating principle applies: AI answers questions; humans make promises.

The channels will keep changing. The definition should not. However a conversation starts, by phone, text, email, chat, an AI copilot, or a channel that does not exist yet, engagement is measured by the customer’s meaningful participation, not by the technology that carried it.

---

8. Conclusion

For years, dealerships have asked: how many leads did we receive? Across 1.04 million leads and 2.5 million customer interactions, that question turns out to measure the wrong thing. More than a third of leads never speak, and they buy very little. The leads that engage in a real conversation before any appointment exists convert at well over twice the rate at every stage of the funnel, at least 1.3 times under the strictest pre-appointment sensitivity, and engaged leads account for 85% of the sales. Every deepening of the conversation reads higher on the dial, to one purchase in five once the exchange reaches the appointment stage. Meanwhile the activity metric this data could test directly as a proxy for interest, the email open, carried no lift at all.

A better question is: how many meaningful customer conversations did we create, how fast, how deep, and what did they tell us?

A lead should no longer be defined by its source. It should be defined by its engagement. The dealerships that measure conversations, create them quickly, deepen them deliberately, and act on what customers say inside them will make better decisions, deliver better experiences, and put themselves in the best position to sell more vehicles.

The conversation is the asset. Everything else is simply the channel through which it happens.

---

About Strolid

Strolid is an automotive Business Development Center (BDC) founded in 2014 and headquartered in Salem, New Hampshire. Built by automotive professionals, Strolid provides outsourced sales and service BDC for franchised dealerships across North America, pairing trained human specialists with an AI layer that assists them rather than replacing them. Over more than a decade, its team of roughly 180 specialists has facilitated more than 3.2 million customer appointments (a company-lifetime figure, separate from the totals analyzed in this study) and today handles more than 150,000 dealer conversations, emails, chats, and texts each month. Strolid also created vCons, a now-standardized (IETF) framework for capturing and protecting conversational data, the same conversation records this study is built on. Its guiding principle is that AI answers questions while humans make promises, which is why no customer ever ends a real conversation talking to a machine. Strolid was founded and is led by Vin Micciche.

Strong Process. Solid Results.

---

References

---

Prepared by Strolid. The findings are based on Strolid’s proprietary BDC operational and conversational data as described in the Methodology. External figures are attributed to their original publishers, and readers are encouraged to consult the cited sources directly.

---

1. DAS Technology. “Automotive Dealership Lead Response Study,” presented at NADA Show 2025 (1,700+ dealerships). https://digitalairstrike.com/das-technology-unveils-comprehensive-automotive-dealership-lead-response-study-at-nada-2025/ ↩︎ ↩︎

2. Pied Piper Management Co. “Pied Piper PSI Internet Lead Effectiveness (ILE) Auto Industry Study,” 2025 and 2026 editions. https://www.piedpiperpsi.com/press/ (2026 study: industry average 71, up six points; 62% of dealerships used both call and text follow-up, up from 49%. 2025 study: 19% of dealerships scored below 40, failing to personally respond. Pied Piper’s press page has since reported a further-updated average of 74; confirm the edition cited against the current release before publication.) ↩︎ ↩︎ ↩︎ ↩︎

3. Cox Automotive. “2025 Car Buyer Journey Study.” Cox Automotive, January 2026. https://www.coxautoinc.com/insights/cox-automotive-car-buyer-journey-study-finds-efficiency-digital-tools-and-ai-drive-record-satisfaction/ ↩︎ ↩︎ ↩︎

4. Urban Science with The Harris Poll. Annual automotive dealer and consumer study (6th annual), 2026. Reported by Auto Remarketing. https://www.autoremarketing.com/ar/retail/urban-science-survey-dealers-confident-in-sales-process-wary-of-ai-for-lead-follow-ups/ (Confirm exact release date before publication.) ↩︎ ↩︎ ↩︎

5. Marchex Institute. “New Marchex Institute Study Reveals Phone Calls Can Make or Break the Automotive Customer Journey.” Marchex, November 13, 2018. https://www.businesswire.com/news/home/20181113005296/en/New-Marchex-Institute-Study-Reveals-Phone-Calls ↩︎ ↩︎

6. Oldroyd, James B., Kristina McElheran, and David Elkington. “The Short Life of Online Sales Leads.” Harvard Business Review, March 2011. https://hbr.org/2011/03/the-short-life-of-online-sales-leads ↩︎

7. CarGurus with NielsenIQ. “2025 Consumer Insights Report.” CarGurus, December 2025. https://www.cargurus.com/press/2025_consumer_insights.html ↩︎