To pilot an outsourced BDC, record 30 days of your own numbers first, hand over one narrow scope, write the success measures and the control before the first call, review at days 30, 60 and 90, and decide go or no-go against what you wrote down. Agree at the start how logins, data and phone routing come back.
Why Pilot Instead of Switching
Handing every phone and every lead to an outside team in one move tells you very little if the month goes badly. Was it the vendor, the handoff, the season, the lead mix? You cannot say, because everything changed at once.
A pilot changes one thing and watches it. It costs less, it puts fewer customers at risk, and it gives you a result you can defend to a partner or a dealer principal.
This is a test of a vendor on a defined slice of work. It is not an overhaul of your own department. If the real problem is an in-house team that needs rebuilding, our 90-day BDC transformation plan is the better starting point, and a pilot can run beside it.
Step 1: Thirty Days of Your Own Numbers
Before you speak to a vendor about dates, pull one full month from your phone system and CRM. Without it, day 90 becomes a debate about how things used to feel.
For the scope you plan to test, record:
- Inbound calls offered and answered, by hour of day and day of week.
- Leads received, with the time each arrived and the time of the first human response.
- Contacts, appointments set, appointments shown and units sold (or service appointments booked and arrived).
- The same lines by lead source.
- What the current coverage costs you for that scope.
Use the definitions on our dealership BDC benchmarks page so the baseline and the pilot are measured the same way. If your CRM cannot produce some of these, write that down too. A store that cannot count its shown appointments has found a problem worth fixing before any pilot starts. A step-by-step method for the pull is in how to audit your dealership's lead handling, and the BDC cost calculator totals the cost side and shows how many of the week's 168 hours your current schedule covers.
Step 2: Pick One Scope
Choose the slice where your baseline shows the widest gap and where a clean line can be drawn around the work. The usual candidates:
| Scope | What the vendor takes | Measure it on | Watch for |
|---|---|---|---|
| After-hours | Calls and leads arriving when your team is off | Answer rate and time to first human response in that window; appointments set and shown from it | Morning handback. Who at the store picks up what was started overnight? |
| Overflow | Calls that ring past an agreed number of seconds, any hour | Answer rate in open hours; abandoned calls | Rollover timing set so short that your own team stops answering |
| One lead source | Every lead from a single provider or form | Contact, set, show and sold rates for that source | A source too small to read in 90 days |
| Service scheduling | Inbound service appointment calls | Answer rate; appointments booked and arrived; advisor time freed | Shop capacity rules. An overbooked Monday is a failed test of your scheduler, not of the vendor |
Do not pilot two scopes at once with the same vendor. The shapes these scopes grow into are compared in outsourced BDC models.
Step 3: Write the Success Measures Before the First Call
Put the measures on one page, signed by the store and the vendor, before launch. A measure written afterward will always fit the result.
A workable page has four parts:
- Two or three primary measures tied to the scope. For after-hours, for example: answer rate in the window, time to first human response on leads, and shown appointments from the window.
- The threshold for each. State it against your own baseline ("better than our baseline month by this much") and write the number in. Your baseline sets the bar. A figure borrowed from someone else's store does not.
- Guardrails. Things that must not get worse: customer complaints, show rate on the vendor's appointments, opt-out handling, notes missing from the CRM.
- The definition of an appointment, word for word, so a set appointment means the same thing in both columns.
Agree on the report at the same time. The pilot runs on the same page described in the weekly report an outsourced BDC should send you, with counts you can reproduce from your own CRM.
Step 4: Choose a Control
A pilot without a comparison is a testimonial. There are two practical controls, and neither is perfect.
Same store, prior period
Compare the pilot months with your baseline month. It is simple, and it is the weakest control, because seasons, incentives, inventory and ad spend all move. Note those changes each month, and set the same months of last year beside it as a second reference.
Split by source or by window
The strongest option when volume allows. The vendor takes one lead source while your team keeps a comparable one, or the vendor takes after-hours while your team keeps the day. Both run in the same weeks under the same market conditions. Be honest about whether the two halves are really comparable: an OEM lead and a third-party lead are different animals.
Whichever control you use, report counts beside rates. At small volumes a rate swings wildly on a handful of appointments, and "9 shown of 14 set" tells a manager more than a percentage does.
Checkpoints at Day 30, 60 and 90
Day 30: is it working mechanically? Calls reach the vendor on the agreed rule. Every contact is in your CRM under a named agent. Recordings open. Transfers are picked up. Corrections from the first weeks are in the playbook. Do not judge sold results yet.
Day 60: is it trending? You now have enough weeks to compare with the baseline on the activity measures: answer rate, response time, contacts and appointments set. Shown results from the first month are in. This is the point to raise a concern formally, in writing, with a date for the fix. It is also when you check your own side: are the vendor's appointments shown the same attention on the floor as your own?
Day 90: the full chain. Set, shown and sold (or booked and arrived) against the baseline and against the control, plus what you paid. Work out cost per shown appointment for the pilot scope and compare it with what the same scope cost you before.
The Go or No-Go Decision, and How to Unwind
Hold the day-90 meeting with the page you signed in Step 3 on the table. There are four honest outcomes:
- Go. Primary measures met, guardrails intact. Keep the scope and decide whether to add another.
- Go, narrower. One part worked (say, after-hours calls) and another did not (after-hours internet leads). Keep what worked.
- Extend. Volume was too low to read. Extend once, for a fixed period, with the same measures.
- No-go. Measures missed, or a guardrail broke. End it. If the failure was on the store's side, such as transfers nobody answered, say so and fix that before testing anyone else.
The unwind should be written before launch, as part of the outsourced BDC contract terms, and it has five parts: phone routing returned to its earlier rule on a set date; every vendor login closed that day (see giving a BDC vendor CRM access); open leads and pending appointments handed back with notes; recordings and any exported data returned or deleted, confirmed in writing; and the vendor's record of do-not-call requests delivered to you.
That last item is easy to forget and it matters. Under the FCC's rules, "a do-not-call request must be honored for 5 years from the time the request is made," and where a third party records the requests, the entity on whose behalf the calls were made is liable for failures to honor them (47 CFR 64.1200(d)(3) and (d)(6), checked October 2026). A request taken by the vendor's agent in week six still binds your store after the vendor is gone. This is general information, not legal advice; confirm the details with your counsel.
Where Pinnacle Fits
We sell an outsourced BDC service, so weigh this section accordingly. BDC On Demand names overflow support, after-hours coverage and event-specific call handling among its engagement shapes, alongside full outsourcing and a hybrid with an in-house team, and those narrow shapes are the natural ones to test first. An engagement starts with a complimentary consultation and a custom plan, and launches with call recordings, performance metrics and regular reporting. Pinnacle does not publish pricing or contract length, so whether a 90-day pilot fits is a question to raise in that first conversation, with your baseline in hand. The rest of the decision-stage articles are in the Dealership BDC Guide.
Frequently Asked Questions
Is 90 days long enough to judge an outsourced BDC?
For answering, response time, contacts and appointments set and shown, yes at most stores. For sold results on a small scope, sometimes not, because the counts are low and deals lag appointments. Decide in advance what you will do if the volume is too thin: one fixed extension with the same measures is the usual answer.
Which part of the BDC is safest to outsource first?
The hours or calls nobody is handling now. After-hours and overflow replace a voicemail or a ring-out, so there is little to lose and the baseline is easy to beat or miss clearly. Handing over work your in-house team already does well is a harder test with more internal friction.
Should I tell my in-house BDC team about the pilot?
Yes, before launch. Explain the scope, why it was chosen and what it does not include. An in-house team that learns about a vendor from the call log assumes the worst, and the pilot then measures resentment. Your agents also need to know how to continue a lead the outside team touched first.
What if the pilot fails because of my own store?
Then it did its job. A pilot often shows that transfers go unanswered, that appointments are not greeted, or that shown appointments are never marked in the CRM. Those problems would undermine an in-house team as well. Fix them first, then decide whether to rerun the test.
Summary
A fair pilot has a 30-day baseline, one scope, measures and guardrails written before launch, a control you have been honest about, checkpoints at days 30, 60 and 90, and a decision made against the signed page. Plan the unwind at the start: routing, logins, open leads, recordings, data and do-not-call records all come back to the store. A pilot that exposes your own handoff problems has still paid for itself.
