Reactivation has the cleanest math in automotive marketing. Most vendor ROI claims still won’t survive finance review. The five numbers a CFO will test and the four hidden costs every honest model has to include.
Most dealer-group marketing budgets have a hierarchy that hasn’t shifted in fifteen years. New lead acquisition sits at the top: Meta, Google, third-party providers, the long tail of paid channels. Below it sit agency fees, content production, and brand campaigns. Below that is a maintenance line for the CRM, which most finance teams treat as overhead.
Sitting inside that CRM, almost always unmentioned in the budget conversation, is the largest underused marketing asset most dealer groups own. Tens of thousands of leads acquired across years of paid campaigns, each one already on the P&L and each one once representing real intent. None of them closed.
Lead reactivation is having its moment in automotive. BCG’s October 2025 research on GenAI in automotive identifies a 35% reactivation rate as achievable with the right architecture, and most major dealer-group conversations now include a reactivation track on the roadmap. The case sounds clean in vendor demos. The case that has to survive the second meeting, the one in front of a CFO with a calculator, looks different.
Some of the math is honest. A meaningful share of it is not. What follows is what an honest reactivation ROI model actually looks like, and why it tends to be the strongest financial argument in most automotive marketing budgets even when the headline numbers are smaller than the vendor decks suggest.
The structure of new lead acquisition ROI is fundamentally counterfactual. A finance team is being asked to approve a spend now in exchange for a projected return over a defined window, built on assumed conversion rates, assumed attribution credit, and assumed channel performance. Each assumption is a place where the model can break down under scrutiny. Most ROI decks paper over this by emphasising the headline figure and burying the assumptions in an appendix slide. Most CFO meetings find them anyway.
Dormant lead ROI inverts that structure. The acquisition cost is already on the P&L. The leads exist in the CRM and can be counted. The original engagement context is preserved. The comparison frame is observable: the dealer group’s current paid acquisition CPL across Meta, Google or third-party channels, benchmarked against actual quarterly invoices rather than theoretical projections.
Three forward variables remain in the calculation: the reactivation rate, the close rate on reactivated leads, and the gross profit per closed unit. Three numbers a finance team can defend, anchored to a sunk cost already incurred. That structural simplicity is the reason dormant-lead math holds up under scrutiny better than most other automotive marketing ROI claims, even when the headline figure is materially smaller than competing new-acquisition projections.
For a dealer group already running conversational AI on inbound traffic, layering a reactivation programme onto the same architecture compounds the case. The infrastructure is built. The brand voice is consistent. The CRM integration already exists. The marginal cost of activating dormant leads is meaningfully lower than the marginal cost of acquiring new ones, and the comparison is easier to defend.
Even with that structural advantage, the typical reactivation pitch fails the second meeting. Four issues show up consistently, and a CFO does not need deep expertise in attribution modelling to surface any of them.
The first is conflation of metrics. A 35% reactivation rate means 35% of dormant leads re-engaged with the brand in some measurable way. They opened an email, replied to a message, configured a vehicle, or booked a test drive. It does not mean 35% closed. Models that carry the reactivation rate through to revenue calculations without separating it from the close rate inflate the projected return by an order of magnitude. Any defensible model has to break out reactivation rate, close rate on reactivated leads, and revenue per closed deal as separate variables with separate assumptions.
The second is hypothetical avoidance. Vendor decks routinely cite “equivalent ad spend saved” as a major line item, comparing the cost of reactivating dormant leads against what it would cost to acquire the equivalent volume of new leads at the current paid CPL. The framing is honest only if the dealer group was actually planning to acquire that volume of new leads at that CPL. If the comparison is against a budget that was never going to be approved, the saving is theoretical and the finance team will say so.
The third is benchmark confusion. BCG’s 35% reactivation rate is a directional industry benchmark, anchored in their consulting work and their research on GenAI in automotive. It is not a guarantee. The variance behind that number is significant. CRM data quality, original lead source, time since last contact, original consent terms, and offer relevance all affect what a specific dealer group will actually achieve. A model that uses 35% as a hard input without stating the variance range is taking a benchmark out of context.
The fourth is behavioural revival. A buyer whose lease was expiring in six months was always going to re-enter the market. A buyer whose original interest was tied to a specific life event was going to come back to the brand regardless of AI intervention. The honest model credits the AI layer for the timing acceleration and the matched offer, not for the entire re-engagement. Some share of any reactivation rate is base-rate behaviour that would have occurred without the programme, and a finance team will eventually ask the question.
A defensible reactivation ROI model rests on five inputs, each of which can be defended independently and which together produce a figure that survives finance review.
The first is dormant lead inventory by age cohort. Count and age. The conventional working window in automotive is 12 to 24 months. Older than that, data quality starts to dominate, the original consent record becomes harder to verify, and the achievable reactivation rate drops sharply. The model has to declare which cohort it is running against and exclude older data unless a specific rationale exists for including it.
The second is the reactivation rate, declared against a stated benchmark. BCG’s 35% is the most widely cited directional input and a reasonable starting point. The credible move in finance review is to declare the benchmark, declare the variance range, and where possible run a small pilot on a single cohort before scaling. A typical pilot runs 200 to 500 leads across eight to twelve weeks. Pilot data anchors the assumption to the dealer group’s own CRM and own buyer base rather than to a published industry average.
The third is the close rate on reactivated leads, measured at the close. Reactivated leads typically close at lower rates than warm new leads from active campaigns, with the caveat that the acquisition cost is already incurred. The variance in customer accounts ranges roughly from 4% to 15%, depending on offer relevance, speed of handoff to the sales team, and whether the buyer’s original blocker has actually been removed. The number has to be measured at the close, not at the booked test drive or the qualification call. Measured at the wrong stage, the model breaks down the first time procurement compares it to actual quarterly close numbers.
The fourth is gross profit per closed unit. The dealer group’s own number, front and back-end combined, not an industry average. This is the line where most vendor models cut corners because vendors do not have access to the dealer’s real economics. Any defensible model has to be run with the actual figure or a range the CFO has explicitly approved. Using a generic industry assumption here is the fastest way to lose the finance team.
The fifth is equivalent CPL avoided. The strongest argument in the model, provided the comparison is like-for-like. The honest version adjusts for lead quality differential. Paid acquisition leads on cold inbound are not directly comparable to reactivated leads with prior brand engagement, so a straight dollar-for-dollar comparison overstates the saving. A defensible adjustment is to apply a quality multiplier of 0.7 to 0.9 to the reactivated lead value when computing the equivalent CPL avoided.
Run with all five inputs and their variance ranges declared, the model produces a defensible range rather than a single hero number. That range is what a CFO will sign off on. The single hero number is what loses the deal.
Most reactivation ROI decks include the licence cost of the AI layer and call the cost side done. A finance review will surface four additional cost lines whether the vendor mentions them or not. Better to have them in the model from the start.
• CRM data hygiene. Most dormant lead databases are 30 to 50% decayed by year two through outdated emails, disconnected phone numbers, changed addresses, and lost consent records. A reactivation programme has to budget for an upfront data cleansing phase, typically four to eight weeks of effort and 8 to 15% of first-year programme cost. Without this phase, the achievable reactivation rate is meaningfully overstated.
• Consent and compliance verification. Reactivation outreach is governed by the consent under which leads were originally captured. Under GDPR specifically, the lawful basis for the original processing has to extend to the proposed reactivation channel and the proposed message. Sometimes that consent is still valid for the channel and offer in question. Sometimes a fresh consent flow is required first, which materially changes the volume of leads available for the programme.
• Sales team handoff design. A reactivated lead arrives at a dealer’s BDC with context the sales team needs to understand and act on. The buyer engaged eighteen months ago, something specific has changed, the AI layer has matched a new offer to their stated preferences. Sales teams have to qualify and close that lead differently from cold inbound. The training, escalation playbook design, and ongoing coaching has a real cost, typically 5 to 10% of first-year programme budget.
• Ongoing signal tuning. The signal detection logic improves with iteration: which triggers correlate with reactivation, which messages convert, which cohorts respond to which offers. This is not a set-and-forget cost. Most dealer groups underinvest in this line, which is the most common reason reactivation programmes plateau at month six. The honest model includes a recurring optimisation budget through the contract term.
A reactivation model that omits these four lines reads as a sales document. A model that includes them reads as an investment case. The difference is what survives the procurement gate.
A typical European dealer-group input set looks something like this. Five thousand dormant leads aged 12 to 24 months, acquired at an average €120 CPL, sunk cost of €600,000 already on the P&L. A reactivation rate benchmarked at 35% with stated variance. A close rate on reactivated leads of 8 to 12% measured at the close. Gross profit per closed unit at €1,500, front and back-end combined.
The headline output of that model is 1,750 reactivated leads producing 140 to 210 closed deals, worth €210,000 to €315,000 in incremental gross profit. Equivalent ad spend avoided at €120 CPL across 1,750 leads is another €210,000, which adjusts to roughly €150,000 to €180,000 once a quality multiplier is applied to account for the lead quality differential between cold inbound and reactivated dormant leads.
That is the number where the vendor deck stops. The CFO meeting goes further. Once the four hidden costs are netted out (€30,000 to €50,000 upfront on data hygiene, €15,000 to €30,000 on compliance verification, €25,000 to €40,000 on handoff design, €30,000 to €60,000 annually on ongoing tuning), the defensible net contribution lands at roughly 60 to 75% of the headline figure. Smaller than the vendor deck. Still the cleanest ROI case in the dealer group’s marketing budget.
Onlive’s AutoROI Calculator runs this math against the dealer group’s own inputs: dormant lead count, original CPL, gross profit per unit, and close rate assumption. The output is structured the way finance teams read it: reactivated leads, equivalent ad spend saved, and a cost-versus-revenue view that compares new acquisition directly against reactivation. The calculator is a directional first pass. The four hidden costs above need to be layered in before the number is taken into a CFO meeting. Run honestly, the math holds up line by line.
For dealer groups already running Onlive’s lead reactivation engine, the inputs to the model are observable rather than estimated. The reactivation rate is measured against the actual cohort. The close rate is tracked at the close. The gross profit per unit comes from the dealer’s own DMS. The variance ranges narrow as the programme accumulates data, and the defensible net contribution becomes provable rather than projected.
For dealer groups still evaluating the case, the question worth answering before the vendor decks pile up is what the math actually looks like against their own inputs. A directional first pass through the calculator takes ten minutes. Layering in the four hidden costs takes another conversation with finance. The result is a defensible range rather than a hero number, and a defensible range is what closes the procurement gate.
The dormant leads are already paid for. The conversion infrastructure is already built. The buyers were once interested enough to engage. The math is already in the CRM. The only question is whether the model can survive the second meeting.