A dealer told me last year that retention marketing didn't work for his store. I asked what he'd run. Email, to the full database, monthly, for six months.
That isn't a test of retention marketing. It's a test of email — and a fairly conclusive one, because a meaningful share of his customers hadn't opened an email from anyone in two years.
This is the most common failure mode I see, and it is almost never a copy problem.
The blast-everything trap
The intuitive fix, once one channel underperforms, is to add channels. Send the email and the text and the mailer. Cover the bases.
What that actually does is train people to ignore you everywhere at once. A customer who gets the same message four ways learns something specific: that your communication carries no information about whether it matters. After that, your genuinely important message — the recall, the lease maturity, the one they'd have wanted — arrives pre-ignored.
Channel proliferation without channel discipline makes retention worse, not better. It just makes it worse more expensively.
What Smart Messaging does instead
The mechanism in Bumper Retention is deliberately unglamorous. As the platform refines and segments a store's data, it tracks which medium each individual customer actually engages with, and concentrates future messaging on that channel.
Not which channel performs best on average. Which one this customer responds to.
The effect compounds in a way that's easy to underestimate. In month one you're guessing across four channels. By month six most of your database has told you, through behaviour, where it wants to be reached — and your cost per engaged customer has fallen without a single change to the creative.
Two things make that possible and both are prerequisites, not features:
Segmentation first. The platform segments 100% of the addressable sales, service and unsold-prospect database into micro-targeted campaigns. Delivery and engagement rates are a function of relevance before they're a function of channel; a well-chosen channel carrying a generic message still fails.
Personalization that's real. Every communication carries the customer's name, the vehicle owned or traded, and is tracked behaviourally back to the customer record in our CDP. "Hi [FIRST_NAME]" is not personalization. Knowing that the vehicle is a 2021 they traded a Civic for, and that they declined a brake job in March, is.
The compliance piece, since it's inseparable
Channel choice isn't purely a performance decision. SMS is the highest-engagement channel in most dealer databases and also the one with the most legal exposure, and the rules moved recently: since April 2025, consumers can revoke consent through any reasonable method — a reply text, a voicemail, an email — and businesses must honour it across all channels within 10 business days. Do Not Call requirements, which used to cover voice only, now cover texts as well.
Our bulk texting respects the opt-in status carried in your CRM or DMS rather than a list somebody exported months ago. That's not a compliance feature bolted on the side; it's the only way a channel-selection engine can be trusted to pick SMS at all.
How to tell if you have a channel problem
Two questions:
- What percentage of your database has engaged with anything you've sent in the last twelve months? If it's under a third, adding channels won't help — you have a relevance problem underneath a delivery problem.
- Could you name, for a given customer, which channel they respond to? If the system can't answer that, it isn't learning. It's just sending.
The stores that get retention right are rarely the ones with the best copy. They're the ones that stopped sending four messages to people who read one.
