Triggered direct mail: sending from your CRM like it is email

Triggered direct mail starts with something that happened in the CRM. Maybe a deal changed stage. Maybe a renewal is getting close, product usage crossed a threshold, or a prospect has gone quiet. That event puts the record into a workflow, where it gets checked for eligibility and a usable address before an approved postcard, letter, note, or gift goes to production.

People compare this setup with email, but speed isn't the useful comparison. Control is. Salesforce or HubSpot can hold the audience logic, message, suppression rules, and reporting, while the physical piece still takes days to arrive. Choose triggers that will still matter after printing and postage. If the idea goes stale in 48 hours, it doesn't belong in the mail.

What triggered direct mail automation actually does

A triggered program gets rid of the recurring CSV handoff. The workflow watches for a specific change and, if the record qualifies, creates a mail request. That request needs a campaign ID, format, template version, personalization fields, requested send date, CRM record IDs, and a unique idempotency key. Don't treat that last one as optional. It's what keeps a sync retry from mailing the same person twice.

Most programs pull from three kinds of signals. Stage-change triggers follow sales activity: a booked demo, a quiet deal, or an opportunity moving to closed-lost. Lifecycle triggers work from customer dates and milestones such as onboarding, renewal, expansion, or a change in executive sponsor. Intent triggers are the least tidy. They react to signs of active research, perhaps repeated pricing-page visits, a jump in account-level web activity, or engagement with a high-value content offer.

Then come the eligibility rules. A record can meet the trigger and still fail the send because the address is incomplete, the contact opted out of physical outreach, or another piece went out in the last 30 days. That's normal. A healthy workflow rejects records.

Stage-change sends that match the sales process

Opportunity stages look like easy triggers because the sales team already updates them. Sometimes. The data is usually messier than it appears on a dashboard, so pin down what each stage means, which timestamp starts the clock, and which later event cancels the piece before connecting anything to production.

A demo-booked send only makes sense when the meeting is far enough away. Tomorrow's call gets an email. A meeting ten business days out gives a letter or postcard time to arrive with an agenda or a relevant proof point. If the meeting moves inside the delivery window, cancel the piece. Do the same if the opportunity closes.

"Gone dark" shouldn't be a stage name. Calculate it. A workable definition is an open opportunity with no buyer reply for 21 days, no future meeting, and two completed follow-ups. At that point, a handwritten note from the account owner can refer to the last discussion and offer one useful next step. Keep it useful. This isn't the place for three paragraphs of pitch copy.

Closed-lost needs breathing room. For a loss tagged "timing," "budget," or "no decision," create a task 90 days after the close date. Recheck the contact and make sure no active opportunity exists, then send a short letter about what changed. Recycling the original pitch is lazy, and the recipient will notice.

None of this works without some discipline in the stage data. Require a loss reason on closed-lost opportunities. Expose meeting dates to the workflow. Log meaningful activity types consistently. If every internal note counts as customer activity, the silence trigger won't fire when it should.

Lifecycle sends for onboarding and renewal

Lifecycle automation runs on dates and completed milestones, not sales stages. It's generally easier to govern because you know the customer relationship and the owner.

During onboarding, the mail should help with the next implementation action. A welcome package after signature can introduce the named contacts and include a kickoff checklist. After configuration, a desk reference or role-specific quick-start card may be more useful. The order matters here: the customer success system writes the milestone back to the CRM, then the send begins.

Renewal mail needs enough runway to affect the conversation. Work backward from the contract date and from how the account team actually handles renewals, not how the process map says they handle them. A letter 120 days before renewal can recap agreed outcomes and invite the sponsor to a planning session. At 45 days, a note can acknowledge the relationship and confirm the next meeting. Either piece should be suppressed if there's an unresolved support escalation, a cancellation request, or a signed renewal.

Be selective. Save higher-cost formats for important account tiers, anniversaries, expansion milestones, or adoption thresholds. Before sending a gift, check the recipient's employer policy.

Intent-based sends need a higher threshold

Intent data is noisy at the person level and often anonymous at the account level. It's a score, not evidence that a named buyer asked to hear from you.

One blog visit should never cause a piece of mail. A defensible rule might require activity on three high-intent pages within seven days, a target-account match, an identified buying role, and no open opportunity. Even then, I'd want a review step at first.

Write only what the signal supports. If all you know is that several people at an account researched a topic, talk about that topic and the account's likely operating context. Never write, "We saw you on our pricing page." It exposes the tracking, can damage trust, and may not even be true of the named recipient.

Put high-scoring accounts into a daily queue so an owner can accept, reject, or change the recipient. Automate that choice later, after the team has seen enough false positives to recognize the patterns.

How Salesforce and HubSpot connect to production

The plumbing is basically the same in either CRM:

1. A record meets the entry rule. In Salesforce, that may be a record-triggered Flow on an Opportunity or a scheduled Flow checking inactivity. In HubSpot, it may be a contact, company, deal, or date-centered workflow.
2. The CRM evaluates gates. The workflow checks account tier, country, contact role, consent or suppression status, recent-send history, opportunity status, owner, and required address fields.
3. The address is verified. A postal service standardizes the address and flags missing units or undeliverable records. Expensive packages also require confirmation that the person works there. This guide to address verification for direct mail covers the main controls.
4. The workflow creates a send request. An API or integration passes data to the mail platform. A custom "Mail Send" object can provide an approval queue and audit trail.
5. The platform renders and fulfills. It merges approved fields into a locked template, generates a proof, prints the piece, applies postage, and hands it to the carrier.
6. Status returns to the CRM. Keep requested, printed, mailed, delivered, returned, and responded as separate values when supported. Sales needs the mailed or delivered event, not the enrollment date.

Sincerely is one example of a platform that can accept Salesforce or HubSpot data for trigger-based postcards, letters, handwritten notes, and gifts, then return tracking and response data. The platform matters less than clean identifiers. Keep campaign IDs and CRM record IDs consistent in both systems.

Only approved CRM fields should flow into print. Define fallbacks and character limits. A blank first name needs to switch to a safe greeting, and the records with the longest values belong in your prelaunch proofs (they're usually the ones that break the layout).

Plan for days, not seconds

An email workflow may send within minutes. Mail has production cutoffs, print time, carrier induction, and transit. A simple postcard submitted before cutoff might enter the mail stream in one or two business days. Letters take longer in some workflows. So do handwritten formats and inventory-based gifts, and weekends or holidays can stretch the schedule further. Transit still varies by class, destination, and carrier conditions.

Don't promise one universal delivery date. Set an expectation for each step:

These are planning ranges, not guarantees. Build in a buffer. If a piece has to land before a meeting or renewal, use the full model in how long direct mail takes.

One more thing. Don't create a sales task when production accepts the job. Create it from the mailed event plus a conservative transit offset, or from a delivery scan. A QR code or short URL response is different: route that immediately because the recipient has acted.

Starter trigger recipes and formats

Start with one or two recipes. The event should be reliable, the audience valuable, and the message relevant for at least two weeks. These cover the common revenue motions.

Cap the first launch and keep a holdout group. For example, enroll 200 qualified records, mail 160, and hold back 40 at random. After a fixed window, compare meetings, pipeline, or renewal outcomes. Scans and replies tell you something about the creative; the holdout tells you whether the mail changed results. See measuring direct mail ROI for the full method.

This part trips people up: define failure paths before activation. An invalid address should create a research task or leave the workflow. Errors retry with the same idempotency key. Returned mail suppresses the next send until someone fixes the address. If a campaign is canceled, stop every piece that hasn't entered production.

Start with one dependable signal

Pick one CRM event the team already records well. Connect it to a format that can arrive before the message goes stale, then add address checks, recent-send suppression, a returned status, and a holdout. Run it at low volume until the loop is dependable. Only then add the next trigger.