Follow-up fails before the first message is sent. Not because the sequence is wrong, not because the timing is off, not because the salesperson forgot. It fails because the record that triggered the follow-up is missing the three fields that would make the message relevant. The operator sends a generic reply. The prospect goes cold. The operator blames the channel.

This is a data architecture problem, not a discipline problem. The CRM is only as intelligent as the fields it captures at intake. When those fields are incomplete, inconsistent, or absent entirely, automation sends noise. And noise from a short-term rental operator — asking about availability for dates the prospect already specified, or following up on a property the prospect already rejected — does not just fail to convert. It signals to the prospect that no one is paying attention.

Most STR operators capture name, email, and phone. Those three fields book no one. The fields that determine whether follow-up works are the ones that carry context: intended check-in window, property type interest, group size or use case, and lead source with campaign attribution.

Intended check-in window tells the system when urgency peaks. Group size or use case tells the system which property to feature and what the objection is likely to be. Lead source with campaign attribution tells the system whether this person arrived from an OTA, a direct site, a referral, or a retargeting ad — which tells the system what they already know and what gap the follow-up must close. Without these four, every message is a cold guess wearing a warm name.

Here is what we typically find when we open an STR operator's CRM pipeline: a "New Inquiry" contact record with a name and an email address, a source field that reads "Website" for 90 percent of entries regardless of actual origin, no check-in date, no property preference, and a follow-up sequence that sends the same three emails to every contact on the same three-day cadence regardless of where they are in the decision window.

The sequence may have taken 10 hours to build. It fires correctly. The automation is working. The system is not. Because when a prospect who asked about a four-bedroom lake property for a corporate retreat gets an email about a studio in the city center on day two, the sequence is executing flawlessly against data that was never collected. The operator's tools are functional. The infrastructure beneath them is not.

OTA-sourced leads and direct-site leads are not the same prospect. An Airbnb inquiry has already seen photos, read reviews, and compared three other listings. A direct-site inquiry may be arriving from a Google search with no prior exposure. The follow-up logic for each is different: one is closing, one is educating.

When source attribution collapses into "Website" or is absent entirely, the pipeline cannot split. Every contact gets the same nurture path. The OTA prospect who is 48 hours from booking receives an educational email about what makes the property unique. The cold prospect who needs three touchpoints before they trust the brand receives a closing offer on day one. Neither converts at rate. Both churn. The operator concludes that follow-up does not work for their market. The actual conclusion is that the data model does not support differentiated follow-up.

Before: The operator opens the CRM on Monday. There are 14 new inquiries from the weekend. Eight have no source tag. Six have no check-in date. The follow-up sequence has already fired on all 14 with the same message. Two prospects have already booked elsewhere. Three have not opened anything. The operator cannot tell which three are worth a personal call because there is no field that indicates lead quality or urgency tier.

After: The intake form captures check-in window, group size, and source with campaign ID. A qualification score appends to the record on creation. The follow-up sequence branches at contact creation: OTA-sourced leads go into a closing track; direct-site leads go into an education track. On Monday, the operator opens the CRM and sees seven high-urgency leads flagged for personal outreach — all with check-in windows inside 21 days. The other seven are in automated nurture sequences matched to their actual interest. No leads are contacted with irrelevant content. The operator makes five calls instead of guessing at 14.

A field that is not required at intake will not be populated at scale. This is not a behavioral observation — it is a structural certainty. If your CRM allows a contact to be created without a check-in window, your sales team will create thousands of contacts without a check-in window. The automation will run. The pipeline will look full. The conversion rate will be quiet and low and no one will trace it back to the missing field.

Required fields feel like friction at intake. They are revenue protection at scale. The operator who enforces five required fields on every contact record has an auditable, segmentable, automatable pipeline. The operator who does not has a contact list with a follow-up sequence attached to it.

Operators who respond to inquiries within five minutes convert at roughly 21 percent. Past 60 minutes, that rate falls to approximately 4 percent. But response speed without data quality only gets the operator in the conversation faster. If the first reply is generic — no acknowledgment of dates, property type, or use case — the conversion advantage of the fast response collapses. Speed plus relevance converts. Speed alone does not. Relevance requires fields.

This is the leak that the System Leak Scorecard surfaces in the first pass: not whether the operator has a CRM, not whether they have a follow-up sequence, but whether the data model underneath the sequence can support differentiated, relevant, timely communication at scale. Run the free STR Leak Scorecard to see exactly where your intake architecture is dropping revenue before your sequences ever fire.