Client readiness: the gauntlet, part two

Part one played fair: symptoms pointed at their causes, and clients told you what they knew. Real engagements grant neither. The four cases below are the harder kind, and each is harder in its own way. In one, the loudest symptom is a decoy and the evidence points somewhere else entirely. In another, the cause is something the client did days before calling you and didn't mention, not out of malice but because they suspect what you'll say about it. In a third, nothing is wrong with the email at all, and the test is whether you can say so with money on the table. The last one stacks decay mechanisms the way real domains stack them: a migration, then a silence, then a spike, each harmless-looking on its own.

The format is the same as part one. Each scenario runs three decisions, each decision builds on the last, and the feedback on the options you didn't pick is where much of the teaching lives. Read it even when you're right.

What's being tested isn't recall; every technical fact here appeared earlier in the course. What's being tested is the order you reach for things when a client is watching, and whether the evidence outranks everyone's favorite theory, including yours.

Your client is a twelve-person accounting firm. Their outsourced IT contractor opens the call on their behalf: 'We're on a blocklist. I ran a scan. Get us delisted and the spam problem goes away.'

The blocklist that wasn't

The firm's monthly client newsletter, sent through a mainstream ESP, started landing in spam at Gmail and Microsoft mailboxes about three weeks ago. The IT contractor's screenshot shows the domain flagged on one list you've never heard of; that list's website offers 'express delisting' for a fee. Everything else on his scan came back clean.

Where do you start?

Start the delisting: pay the fee, it's small next to the client's lost business.Agree the listing needs handling, but use the list's free removal form instead of paying.Thank him for the scan, then pull your own evidence: get a copy of a newsletter that landed in spam and read its Authentication-Results header.Open a ticket with the ESP; placement happens on their infrastructure.

Your client is the owner of a commercial cleaning company. 'Our emails started going to spam last week. We haven't changed anything. Can you fix it today?'

Nothing changed

The company sends a monthly newsletter and quote follow-ups from its main domain through a mainstream ESP. The owner is adamant that nothing changed. A first pass shows SPF, DKIM and DMARC present, passing and aligned on a test message you trigger yourself.

The records look fine. What next?

Tell the owner the technical setup is clean, so the problem must be content, and offer a rewrite of the templates.Ask for the ESP's full campaign history for the last thirty days, everything sent, including anything outside the usual newsletter.Check the domain and sending IP against the major blocklists.Set up Google Postmaster Tools and wait for reputation data to accumulate.

Your client runs a specialty coffee subscription. 'Open rates are down forty percent since spring. We're obviously in spam. Our last consultant couldn't fix it. We're hoping you can.'

The clean audit

You run the full audit you'd run on any stranger's domain. SPF, DKIM and DMARC pass and align; DMARC sits at p=quarantine with reports flowing; no listings on any blocklist a real receiver consults; Google Postmaster Tools shows domain reputation High and user-reported spam under 0.1% all year; a placement test to mailboxes you control lands in the inbox at every major provider. The client repeats: opens fell forty percent.

The plumbing checks clean. What do you do with the claim?

Keep digging on placement: buy a commercial seed-list test across dozens of mailbox variants until something shows spam.Stop auditing the mail and start auditing the timeline: when exactly did opens fall, in one step or a slide, in which segments, and what did the business change around that date?Tell them the metric is broken: Apple Mail Privacy Protection made open rates noise years ago.Check Gmail's Promotions tab; the mail is probably inboxing but tabbed.

Your client is an online gift retailer. It's the second week of December. 'Black Friday was a disaster. Our biggest campaign of the year, and half of it bounced or went to spam. We need this fixed before the January sale.'

The December call

The history arrives in fragments. The company switched ESPs in April. Email 'was fine' through spring. They barely sent over the summer ('nothing to announce'). Then in Black Friday week they sent four campaigns to the full list, roughly six times their normal monthly volume. The owner forwards a bounce: a 4xx deferral whose text says the receiver detected an unusual rate of mail from the sender and is rate-limiting it.

Where do you start, with January six weeks out?

An evidence sweep before any change: full health check on the domain, headers from a spam-landing message, the bounce texts sorted 4xx versus 5xx, Postmaster Tools history back to April, and the DNS zone as it stands today.The bounce names rate limiting, so throttle the send speed and re-send the failed portion.The ESP switch is the obvious suspect: open a migration-review ticket and let the new ESP audit its own setup.Write off Black Friday and plan January as a fresh, carefully warmed campaign.

The readiness signal

Score yourself across both parts of the gauntlet, and score the first attempt, not the replay. Every decision landed in one of three places: you picked the strongest move outright, you picked a defensible move and saw why the stronger one was stronger the moment you read the feedback, or the feedback had to talk you out of something.

If most decisions landed in the first two, take that record to the graduation checklist in the final lesson and check yourself against it line by honest line. The gauntlet is the evidence for the judgment items on that list: evidence before theory, cause before symptom, fixes in sequence, and the nerve to say 'this isn't an email problem' when it isn't. That last one deserves its own sentence. Nobody fails this course by not knowing what a selector is; the drills caught that long ago. The way to fail it is to sell a fix where nothing is broken, and scenario three exists because that temptation will arrive in real life with a real number attached.

If whole scenarios went wrong at step one, that's information too, and cheaper now than later. Misses cluster: jumping to fixes before evidence points back at Part 6, trusting the client's framing points at the triage tree, misordered repairs point at the fixing-in-order lesson. The cluster names what to sit with again before you charge anyone.

And one plain statement to close the exam on: the gauntlet is practice. It's the best rehearsal we could build from the failure classes our checks catch on real domains, and rehearsal is what it is. Your first paying client is the real part two: their symptoms will arrive described worse, their timeline will surface slower, nobody prints feedback under the options, and the invoice has your name on it. Passing here means ready to start, not finished; the distance between those two is the first few engagements, done carefully.

Graduation is next: the full competence checklist, self-assessed against every checkpoint and this gauntlet, plus the habit that keeps a practitioner current after the course stops updating them.