A Marketplace alert is not a buy decision. It is a ticket that says "this listing cleared your filters at ingest time." The buyers who stay fast treat that ticket like a triage queue: kill junk in seconds, then spend attention only on listings worth a human message.
If every ping opens a ten-minute rabbit hole, your monitor is working and your process is not.
Why triage beats "open everything"
Good filters still let imperfect listings through. Sellers typo prices. Titles omit years. "Camry" shows up next to "parts only" in the body after you already got the alert. Recency can look fine while the car is three cities farther than you will drive tonight.
Triage is the discipline that absorbs that noise without widening your price cap again. You are not trying to be thorough on every row. You are trying to protect the minutes that should go to messaging real sellers.
The two-minute ritual
Run the same steps in the same order every time. Stop as soon as a listing fails.
Minute 0:00 to 0:20: read the receipt
Open the alert and read the match reasons before the photos. Confirm which checks passed: make and model, price, year, excluded terms, recency. If a reason looks wrong for your buy box, fix the filter later. Do not argue with the listing first.
Minute 0:20 to 0:50: kill on hard fails
Open the Facebook listing. Reject immediately if any of these are true:
- Price is a clear typo or outside the range you actually pay
- The vehicle is parts, wrecked, no title, or "bill of sale only" when you need a clean title
- Location is outside the radius you will travel today
- Photos are stock images, repeated across other ads, or missing entirely
- The seller profile looks brand new and the copy is generic pressure language
Hard fails do not get a message. They get a note in your head for a filter tweak if the same pattern repeats.
Minute 0:50 to 1:20: soft checks
If it survived the kill list, scan for soft risk:
- Does the year in the title match the description?
- Are miles, trim, or title status mentioned at all?
- Is the asking price far enough under local comps that you should expect a scam or a catch?
- Has the listing been edited into a different car than the alert implied?
Soft risk does not always mean walk away. It means your first message must ask a pointed question instead of "Is this available?"
Minute 1:20 to 2:00: message or defer
Only two outcomes remain:
- Message now with one specific opener (year confirmed, asking about title status, requesting VIN, offering a daylight inspection window)
- Defer if you cannot act today (wrong city for this week, budget already deployed). Deferred means closed for now, not parked in a guilt pile
If you cannot decide in two minutes, the listing is probably soft-risk heavy. Ask one clarifying question or move on. Do not start a research project mid-alert.
What to write down after a bad streak
When three alerts in a row are junk for the same reason, change the monitor, not your standards.
Examples:
- Three "parts" cars → add excluded terms
- Three out-of-area listings → shrink cities or radius
- Three yearless wrecks → raise the year floor or require clearer titles in your soft checks
- Three dealer dumps → exclude "dealer," "credit," or "down payment" if those words keep appearing
Filter edits affect new ingests. Old alerts keep the reasons they had when they fired. That is expected. Tighten forward.
What Crawlbench is for in this ritual
Crawlbench's job ends at a filtered alert with reasons. The vehicle monitoring flow and features overview cover how scheduled checks and the match gate work. Fan-out runs every 10 minutes, with city coverage rotating across anchors.
Your job starts at triage. The monitor should shrink the queue. You still decide who gets a message.
Closing rule
Two minutes per alert is enough when the first twenty seconds are spent on the receipt, not the photo carousel. Kill fast, message specific, tighten filters when the same junk repeats.