Lead Nurturing Should Decide What Happens Next

Lead nurturing is the process of keeping an identified prospect usefully engaged when an immediate sales conversation is not yet the right action. The practical goal is not to deliver a predetermined number of emails. It is to decide what should happen next as new evidence appears: continue helping, change the message, pause, hand the prospect to sales, or stop.

That definition creates a useful test. If a campaign cannot explain why a person entered, what evidence would change their path, what would prevent the next action, and how the journey ends, it is an email sequence—not yet a complete nurture program.

Oracle similarly describes nurturing as cultivating leads that are not ready to buy and says communications should reflect both the prospect’s profile and position in the buying process. Crucially, its guidance treats nurturing as a workflow whose steps have distinct objectives, rather than as email alone (Oracle).

Start with one transition, not a content calendar#

Before selecting content or cadence, name the state change the program is supposed to produce. For example:

Move eligible prospects who are not ready for sales either to an accepted sales review or to a recorded non-sales outcome.

This is more useful than “warm leads” or “increase engagement.” It identifies a starting population, a destination, and an observable result. It also makes non-sales outcomes legitimate. A prospect who is ineligible, opts out, lacks required information, or remains inactive should not have to become a sales lead for the program to reach a valid conclusion.

Write the operating contract before building automation:

This one-page contract is the working specification. The automation canvas implements it; it should not be the first place the logic becomes visible.

Turn prospect activity into explicit decisions#

A useful branch connects evidence to a materially different next action. Write each branch in this form:

current state + accepted evidence + guard
→ next state + action + reason

For example:

The “no” branch deserves careful wording. No tracked event within a window means only that the program observed no qualifying event. It does not establish that the buyer has no interest. Treating absence of telemetry as a fact about intent makes the workflow look more certain than its evidence.

Keep behavior and current data separate as well. Salesforce’s workflow documentation distinguishes triggers, which wait for prospect actions during a specified period, from rules, which inspect current data when the prospect reaches that step. Actions then do something, and end steps terminate a path (Salesforce Trailhead). The product terminology is specific to Salesforce, but the design question is portable: are you waiting to observe an event, or checking what is true now?

If two branches lead to the same state, owner, and treatment, remove the split. It adds reporting complexity without changing the decision.

Use waits as observation windows#

There is no useful campaign length independent of the decision being observed. Instead of asking how many days a nurture should run, ask how long a particular branch needs to wait for evidence that would alter the next action.

For every wait, specify:

• the event or condition being observed;
• when the window opens and closes;
• what happens if evidence arrives early;
• what happens if nothing is observed;
• which eligibility and suppression checks run before the next action; and
• whether a pause freezes, restarts, or ends the timer in the selected platform.

Do not collapse all forms of inactivity into one status. Adobe Marketo, for example, distinguishes a person’s Normal or Paused cadence from Exhausted, which means the person has received every item in a stream (Adobe). Exhausting content does not prove readiness, disinterest, or permission. It only reports that no unused content remains in that stream.

Recheck permission and commercial state before every action#

Eligibility at enrollment can become stale. A prospect may unsubscribe, encounter a hard delivery failure, enter an active opportunity, become a customer, or receive direct sales outreach while waiting. The program therefore needs a send-time gate, not merely an entry filter.

Evaluate vetoes in a declared order:

1. Resolve the person and destination.
2. Check whether the purpose, topic, sender, and channel are currently permitted.
3. Check delivery blocks and complaints.
4. Check whether the current customer, opportunity, or sales-ownership state makes the treatment inappropriate.
5. Apply frequency, priority, quiet-hour, holdout, and incident controls.
6. Confirm that the journey is still active and has an eligible next action.

Record the reason and scope of every veto. suppressed, paused, exited, and exhausted are not interchangeable. Suppression forbids an action within a defined scope; pause delays execution; exit ends this journey; exhaustion means that a stream has no remaining content.

Applicable privacy rules vary by jurisdiction and channel, so the responsible privacy owner must define the policy. As one concrete regulatory example, the UK Information Commissioner’s Office says an organization should retain the minimum information needed on a suppression or do-not-contact list after an objection or opt-out, rather than simply deleting the contact and risking future re-import. Its guidance also notes that an opt-out may apply to a specific channel or activity, while an objection can be broader (ICO).

Make sales handoff a transfer with a receipt#

A score change, CRM update, task, or webhook can initiate a handoff. None proves that sales accepted responsibility.

Send a compact handoff package containing:

• prospect and account identity;
• current lifecycle and nurture state;
• the fit and readiness evidence that crossed the gate;
• evidence timestamps and sources;
• the problem or content context relevant to the first conversation;
• preference and channel constraints;
• assigned owner and response expectation; and
• a correlation ID connecting the request to the response.

Require the receiving process to return a controlled disposition such as accepted, rejected_with_reason, duplicate, or needs_context. Until that receipt arrives, keep the record in handoff_pending. This exposes stuck transfers and prevents marketing from reporting a task creation as sales acceptance.

Define the return path at the same time. A timing rejection may recycle into a bounded nurture path. A fit rejection may exit this program. A duplicate should resolve to the authoritative record. A request for context should return to an owner who can supply it. The reason code determines the next action; “rejected” alone does not.

Measure whether the decisions worked#

Email delivery, opens, and clicks can help diagnose content and mechanics. They do not show whether the program produced the promised transition. Report a versioned cohort of prospects admitted under the same eligibility and entry rules, then measure the state changes that matter:

Keep the numerator, denominator, and program version visible. If eligibility, branches, content, or the sales agreement changes, start a new version or segment the report. Otherwise, a change in results cannot be separated from a change in the operating definition.

Before broad enrollment, trace at least four representative records end to end: no observed behavior, an early high-intent event, a suppression arriving during a wait, and a handoff returned by sales. For each trace, confirm the source facts, branch decision, veto result, emitted action, receiving-system record, and terminal reason.

The launch gate is simple: another operator should be able to explain why this prospect received—or did not receive—this action at this time. When the workflow can do that, lead nurturing becomes a controlled way to choose the next useful step. Until it can, keep the audience and automation bounded; increasing volume will only scale the ambiguity.