The 06:00 briefing: how I run 14+ brands without opening 14 dashboards

The most useful thing I've built for an ecommerce business is also the most boring, and I mean that as a compliment. It's a briefing that lands at 06:00, before I'm properly awake, and it reads across the whole business — Shopify, Meta, Google Ads, Klaviyo, GA4, Search Console — so that by the time I open the laptop the overnight mess has already been turned into a single packet I can actually read. No demo where a robot does something clever and everyone nods along. It just sits off to the side and catches the stuff that quietly costs you money.

I know it catches that stuff because it's bitten me. I've had a broad-match test on our own SFT account wander off and start spending on terms it had no business spending on, and I've had server-side events quietly stop firing so the numbers looked fine right up until they didn't. You don't forget those, and they're a big part of why the briefing exists in the first place — it watches for the things I've already been burned by.

The business should be read before the operator arrives

Every morning, before you're up, the system pulls the numbers and checks the signals across each account, and it's answering a few questions for you — what changed, what broke, what needs a decision today — and one more that usually gets skipped, which is what's actually fine and can stop asking for your attention. A good briefing shouldn't just manufacture alarms; if nothing needs you, the system should say so and let you drink your coffee.

In my own portfolio system that daily layer reads across 14-plus brands and scores what needs attention before morning, and the point of all of it isn't to replace the judgement call — it's to have the judgement call teed up and waiting when you sit down, instead of you reconstructing last night from scratch.

What goes into the packet

• revenue, spend and efficiency movement
• account delivery issues
• tracking weirdness — the server-side events that quietly stop firing
• creative fatigue
• negative keyword waste
• budget pacing
• email and customer signals
• decisions carried forward from yesterday

Then it gets written in a way a human can actually use — not a wall of numbers, but a verdict with the evidence behind it, a recommended next action, a confidence level, and the caveats where the data's only partial. It shouldn't dump a toolkit on the floor and ask you to assemble your own morning out of it; it should point at the next sensible step and tell you why.

Why timing matters

A 06:00 briefing changes the day because it lands before the day fragments — before the calls, the Slacks, the small fires that turn a clear head into a reactive one — so you're not diagnosing last night's performance in the ten minutes between meetings, you start with the shape of the work already visible. For an agency that's the bit I'd push hardest: the client conversation starts from evidence instead of "we're keeping an eye on it", which is usually code for the dashboard existing somewhere that nobody's looked at since Tuesday.

It also leaves a memory trail. Every day leaves a record, so the decisions compound and the weirdness doesn't get forgotten, and next time something odd shows up the system already knows whether it mattered last time.

The human still decides

This part matters to me. I don't want a system that wakes up and starts changing client accounts on its own. The briefing surfaces the decision; the human makes it.

Over time some of the guardrails can become more automatic — but only where the action is narrow, reversible and well understood, like a negative keyword suggestion or a pacing warning. Not the broad strategic calls. Those stay with you.

A demo can impress you in a meeting. A morning system has to behave on the day the numbers get weird, and the test of a good one isn't whether it makes you feel replaced — it's whether you walk into the day already briefed.