Sign up | Read online

Hey, it’s Jochem. Welcome back to the Ecom Handbook #13.

This time number 13 is a lucky number! hahaha

Quick question: is there anything I can do better with these newsletters?
Just hit reply and tell me. I read every message and it directly shapes what I write next, my goal is to make each edition more useful than the last to help you grow your online store with actionable tips.

-Jochem (@iamjochems)

ROAS drops don’t start in the ad account.

They show up there, but they rarely start there.

Most teams respond the same way:

Pause ads.
Cut budgets.
Launch new creatives.
Blame the algorithm.

Movement feels productive.

It usually isn’t.

A drop in ROAS is a signal.
Your job isn’t to panic.
Your job is to diagnose.

Let’s walk through how professional operators do it.

There are only four real categories:

If you don’t know which one you’re dealing with, every action is guesswork. Most brands treat all four the same.

Let me tell you, they aren’t.

Let’s break them down.

Before touching budgets, ask:

Did reality change, or just reporting?

Check:

Look at:

If platform ROAS drops 25% but MER is flat, you don’t have a performance problem. You have a measurement problem.

Different response entirely.

If traffic is stable but conversion rate falls, the issue isn’t acquisition.

It’s friction.

Check:

Also check seasonality.

A 3.2% CVR dropping to 2.4%
with stable CPC
will crush ROAS.

Ad managers won’t fix that, operators go to the funnel.

Sometimes the algorithm expands, and expansion lowers average intent.

Signals:

That can be healthy.

Early-stage scale often compresses ROAS before revenue catches up.

Ask:

Are we buying worse traffic? Or broader traffic?

If new customer acquisition improves and blended revenue rises, lower platform ROAS may be acceptable. Context matters more than screenshots.

This one hurts the most.

Same traffic.
Same CVR.
Same structure.

But:

ROAS falls because unit economics changed.

Common causes:

If margin compresses 10%, ROAS must rise to maintain the same contribution. Most teams never adjust targets when economics change. They just demand “performance recovery.”

Math doesn’t care..

If you skip classification, you misdiagnose. If you misdiagnose, you destabilize. And instability compounds.

Pull 8–12 weeks of data, look at:

Ask:

Is this a platform issue or a business issue? If blended performance is stable,
don’t torch campaigns.

Compare:

Last 14 days vs Prior 14 days vs Same period last year (if relevant)

Identify what actually moved:

Only one or two metrics usually shift, find the lever before pulling ten.

Force it into one bucket:

Measurement
Conversion
Traffic
Economics

If you can’t clearly classify it, you don’t understand it yet. Keep digging!

Measurement issue?
Fix tracking. Don’t cut spend.

Conversion issue?
Fix offer, UX, or inventory.

Traffic expansion?
Adjust constraints slowly. Tighten creative angle before structure.

Economic compression?
Recalculate target CAC and allowable ROAS, then realign budgets.

Precision beats aggression.

Nothing amplifies a dip like chaos.

When performance drops, teams increase change velocity.

That resets learning.
Which increases volatility.
Which increases panic.

Instead:

Make one structural adjustment.
Wait.
Evaluate directionally.

Stability is underrated alpha.

Use AI as a diagnostic analyst.

Prompt:

Paste:

If AI struggles to classify it clearly, your data story isn’t clean. That’s insight. AI is strong at pattern recognition, use it to reduce emotional bias.

Daasity centralizes Shopify, ad platforms, and financial data so you can:

Best for: distinguishing traffic problems from margin problems.

Rockerbox helps you:

Best for: understanding whether the drop is real or a channel mix artifact.

AI trends shaping e-commerce strategies

A new “Agentic Commerce Shift Report 2026” outlines key execution shifts e-commerce leaders are using to boost conversion and profitability as AI becomes central to online retail operations.

Read it here

Why AI agents are reaching checkout without clear rules

AI agents are moving from answering questions to taking actions (such as checkout) inside browsers, business software and automation platforms, but a new research index finds that the rules are lagging their commercial use.

Read it here

Open a blank doc.

Title it:

“Why Our ROAS Dropped.”

Force yourself to write one paragraph answering:

Which of the four categories caused it?

If you can’t confidently pick one, you’re reacting - not diagnosing.

Then write: What is the single highest-leverage correction? Execute it.

Important: Wait long enough to measure signal direction.

Because most ROAS drops don’t require heroics, they require clarity.

And clarity compounds.