You're probably here because it's Friday, revenue is down 30% week-over-week, and you can't explain why. Or it's Monday morning, you've just opened Shopify admin for the first time in three days, and something obviously broke while you weren't looking.

Whatever triggered the search, the feeling is the same. Something's wrong, you don't know what, and your instinct is to start changing things fast. Brief new ad creative. Email your agency. Turn a campaign off. Switch a flow back on. Tweak a price.

That instinct is the most expensive mistake founders make in the first 24 hours of a revenue drop. Not because action is wrong, but because action without diagnosis almost always fixes the wrong thing, burns another week, and sometimes makes the original drop worse.

This post is the diagnostic framework I wish every DTC founder had taped next to their monitor. Five ordered steps, a first-30-minutes checklist, a table that maps common symptoms to their actual root causes, and a clear rule for when to pause spend versus sit with it. The goal is to turn a panicked Friday into a structured investigation that ends with an accurate answer, not a faster wrong one.

#Why Most Founders Misdiagnose Revenue Drops

Five patterns come up repeatedly. If you recognise yourself in any of them, that's fine. Most founders have done all five at some point.

Acting before investigating. The revenue drop is real, the founder is anxious, and within two hours they've briefed three new creative concepts, paused a campaign, and scheduled an email blast. None of it is based on a diagnosis. It's based on the need to feel productive. Half the time, the drop was already correcting itself by the time the "fix" went live.

Blaming the most visible channel. Paid ads get blamed for almost every revenue drop, whether or not they're responsible. They're visible, they cost money, and they feel controllable. The real cause is often something quieter: a broken flow, a checkout bug, a refund batch, an inventory gap. Ads are the first place founders look because they're the first place founders look, not because they're the most likely culprit.

Overcorrecting on noise. Weekly revenue is noisy. A 15% drop in a single week can be a genuine problem or it can be a Tuesday that happened to follow a big Monday. Founders who react to every weekly dip burn out on false alarms and start missing the real signals.

Missing attribution artifacts. Sometimes revenue isn't down, it just looks down because a reporting tool changed how it counts something. iOS updates, cookie changes, a new UTM scheme, a broken pixel: all of these can show up as "revenue dropped" in a dashboard while the actual money in the bank is unchanged. Diagnose the data before you diagnose the business.

Fixing the symptom, not the cause. A founder sees conversion rate drop, assumes the product pages need work, and briefs a landing page test. Two weeks later, the test is inconclusive and conversion is still down. The real cause was that paid traffic shifted toward a new interest audience that doesn't convert as well, and the product pages were fine. The symptom was conversion, but the cause was traffic quality.

Good diagnosis is boring and sequential. Bad diagnosis is fast and confident. You want the boring version.

#The Diagnostic Framework

The framework is five ordered steps. The order matters. Each step narrows the possible causes so the next step is faster. Skip one, and you'll spend hours investigating the wrong layer.

#Step 1: Verify the Drop Is Real

Before you investigate anything, confirm that revenue actually dropped in the bank, not just in a dashboard.

Go to Shopify admin. Look at gross and net revenue for the period in question. Compare it to the same period last week and the same period last month. If Shopify's own numbers agree that revenue is down, the drop is real. If Shopify's numbers look fine but an external dashboard (Triple Whale, Northbeam, GA4) shows a drop, the drop is an attribution or reporting artifact, not a business problem.

Three specific artifacts come up often enough to check for every time:

• A tracking pixel broke or a UTM scheme changed, so attributed revenue looks lower even though Shopify revenue is flat
• A reporting tool is comparing unequal periods (a seven-day window that includes a holiday against one that doesn't)
• A data source is lagging by 24 to 48 hours, so today's number looks low because it isn't finished loading

If any of these apply, stop. There's no revenue drop, there's a data problem. Fix the data.

#Step 2: Split New Customer vs. Repeat Revenue

Once you've confirmed the drop is real, this is the single most useful cut you can make.

In Shopify, pull new customer revenue and returning customer revenue for the drop period and the comparable prior period. One of three things will be true.

If new customer revenue is down and repeat revenue is stable, it's an acquisition problem. The issue is somewhere in your paid channels, organic traffic, or top-of-funnel conversion. You don't need to look at email flows or repeat purchase behaviour. Eliminate them from the investigation.

If repeat revenue is down and new customer revenue is stable, it's a retention problem. Something in your post-purchase experience (email flow, SMS, subscription, returning-customer promo) has changed. You don't need to look at ad accounts. Eliminate them.

If both are down, it's either a site-wide issue (checkout broken, product unavailable, storefront bug) or a category-wide shift (seasonal, market-wide, competitor launch). The investigation is broader, but the next step will narrow it.

This single split eliminates half the possibilities in under five minutes. Most founders skip it and investigate everywhere at once.

#Step 3: Identify the Channel

Now pull revenue by source for the drop period versus the comparable prior period. Look at the mix, not just the absolute numbers.

You're looking for one of two patterns. Either one channel dropped dramatically and pulled the total down, or every channel dropped proportionally, which suggests a site-level issue rather than a channel-level one.

If one channel dropped, go there and keep narrowing. If it's paid, look at spend, CPM, and CTR. If it's email, look at flow sends and open rates. If it's organic, look at search rankings and referring sites.

If every channel dropped proportionally, you're almost certainly dealing with a site or product problem, not a marketing problem. Skip ahead to Step 4 and look at conversion rate.

#Step 4: Identify the Funnel Stage

For the channel (or channels) that dropped, walk the funnel from top to bottom.

• Did sessions drop? That's a traffic problem
• Did sessions hold but add-to-cart drop? That's a product page or product availability problem
• Did add-to-cart hold but checkout initiation drop? That's a cart or shipping display problem
• Did checkout initiation hold but conversion drop? That's a checkout or payment problem

Most revenue drops live in one of these four buckets, and the bucket tells you where to look next. A traffic drop means the issue is upstream (ad account, search ranking, referral source). A conversion drop means the issue is on-site (bug, friction, offer change).

The bucket you land in should match the channel from Step 3. If it doesn't, you've either split the data wrong or there are two things going on at once.

#Step 5: Find the Change Event

Every sudden revenue drop has a change event behind it. Something changed in the last 14 days, usually the last 7, often the last 72 hours. Your job in Step 5 is to find it.

Make a list of everything that changed recently:

• New ads launched, old ads paused
• Changes to audiences, budgets, or campaign structures
• Theme or code deployments (even tiny ones)
• New apps installed or removed
• Product launches, restocks, or price changes
• Email flow edits, flow pauses, or new segments
• Shipping policy or rate changes
• Homepage or PDP updates
• Inventory changes (items going out of stock)
• Refund batches or chargeback spikes

Cross-reference the list against the channel and funnel stage from Steps 3 and 4. One of those changes is usually the cause. If nothing on your list matches, widen the window (look back 21 days) or look at external events: platform changes, competitor moves, seasonal shifts.

The change event is often embarrassingly small. A checkout setting toggled. A flow paused "temporarily" three weeks ago. A pixel that stopped firing after a theme update. The smaller and older the change, the harder it is to find, and the longer it usually sits before anyone notices.

#The First 30 Minutes: What to Check

Here's the condensed checklist for when something's wrong and you've got half an hour before your next call. Run these in order.

That's ten checks. If done efficiently, you can run them all in 30 minutes and walk out with a specific hypothesis about what went wrong. You don't need to solve the problem in 30 minutes. You need to narrow the possibilities enough that the next two hours of work are actually useful.

#The First 48 Hours: Deeper Investigation

If the 30-minute pass didn't produce a clear answer, the next 48 hours are for the slower checks.

Look at historical drops of similar size. Open your past 12 months of weekly revenue data. Was there a drop this size before? When? What caused it? If drops of this magnitude are normal (say, around holidays, promotions, or post-launch dips), you may be reacting to noise.

Look at category and competitor signals. If you're in a category that moves with news cycles, weather, or macro events, check whether peers are seeing the same pattern. If they are, the drop is category-wide, not a you-problem. Adjust your response accordingly.

Look at audience and creative fatigue. If paid performance dropped and nothing else changed, the creative may have fatigued or the audience may have saturated. CPM rising with CTR falling is the classic pattern. If you see it, the fix is new creative, not spend changes.

Look at attribution shifts. If your reporting tool changed, or iOS or Chrome pushed an update, or a pixel was modified, revenue attribution can shift even when the business didn't. Compare attributed revenue in your reporting tool to actual Shopify revenue for the same period. Any meaningful gap is an attribution artifact.

Look at the change log. If nothing on your 14-day list matches, widen to 30 days. Some drops are delayed reactions. A flow paused three weeks ago doesn't hurt immediately because existing subscribers still convert. It starts hurting in week three when the new subscribers would have started converting.

#When It's Not Actually a Drop

Sometimes the drop isn't a drop. Before you start fixing things, rule out these common false alarms.

Day-of-week effects. If your best days are Monday and Tuesday and you're looking at a Saturday, the "drop" may be normal seasonality within the week. Always compare like days, not running totals.

Uneven comparison periods. Comparing this week (seven days) to last week (seven days that happened to include a holiday or a promo day) will always show a drop. Check whether the comparison periods are actually comparable.

Normal volatility. DTC brands under $500K per month see 20% week-over-week swings regularly. Brands doing $50K per month can see 40% swings that mean nothing. The smaller the business, the noisier the weekly data. Set a threshold below which you don't investigate.

Attribution changes. Already covered, but worth repeating. Your reporting tool may have changed how it counts something. Always validate against Shopify's own numbers before you treat a drop as real.

A single outlier day. One bad day doesn't make a trend. If yesterday was weak but the week before yesterday was fine, and the day before yesterday was fine, wait 24 to 48 hours before acting. Single-day drops often revert on their own.

A good rule: if the drop is under 15% week-over-week, revenue is above $20K per week, and there's no obvious change event, give it another day before you act.

#Common Root Causes and Their Fingerprints

Once you've run the diagnostic, you're looking for the pattern that matches your data. Here are the most common causes and what they look like.

Match your symptoms to the fingerprint. Usually one row fits cleanly. If two rows fit, you have two problems at once, which happens more often than founders expect.

#When to Act, When to Wait

Once you have a hypothesis, the next question is whether to act immediately or gather more data.

Act immediately if:

• Something is clearly broken (checkout error, flow paused, pixel down, site bug)
• Spend is obviously misallocated (an ad with 0.3x ROAS running three days in a row)
• A flow stopped sending and the fix takes 10 minutes

Wait 24 to 48 hours if:

• The drop is within your normal volatility range
• You don't yet have a clear hypothesis
• The most likely cause is external (market, competitor, seasonal) and not something you can fix
• Acting would mean a major change (pausing a working campaign, pulling a product, major email reshuffle) based on incomplete diagnosis

The cost of waiting 48 hours is small. The cost of pausing a working campaign based on a bad diagnosis is large. When in doubt, wait.

One exception: if the drop is clearly accelerating and you can identify a specific lever to pull, pull it. A pattern that gets worse every day is different from a pattern that stabilises.

#A Worked Example

Here's how this plays out in practice.

You're running a $400K-per-month supplements brand. Monday morning you open Shopify and see revenue down 28% week-over-week. Your first instinct is that Meta Ads are underperforming and you should brief new creative.

Instead, you run the framework.

Step 1: Verify. Shopify admin confirms gross revenue is down 28%. The drop is real, not an attribution artifact.

Step 2: New vs repeat. New customer revenue is down 12%. Repeat revenue is down 41%. The problem is mostly retention, not acquisition. You just eliminated half the possibilities. Ads are not the primary cause.

Step 3: Channel. Email revenue is down 58%. Every other channel is roughly flat. The drop is concentrated in email.

Step 4: Funnel stage. Email sessions are down, not email conversion rate. Something is reducing email traffic, not email effectiveness.

Step 5: Change event. You check Klaviyo. Your post-purchase flow has been paused for nine days. Someone edited a segment two weeks ago and the flow got caught in a deactivation. The last two weeks of new customers never received the follow-up emails that drive the second purchase.

The fix: unpause the flow. Rebuild the affected segment. Email the last two weeks of customers manually with a catch-up message. Total work: 45 minutes.

If you had started with "let's brief new Meta creative," you would have spent a week of spend on a problem that didn't exist and left the real issue running.

The framework is not about being clever. It's about not being wrong.

#Building a System That Catches Drops Earlier

The best revenue drop is the one you catch before it compounds. Most of the damage from a drop isn't the first week, it's the three weeks after the first week that you didn't notice.

A few habits help.

Weekly review cadence. A 30-minute weekly growth review is the single most effective system for catching drops early. You're scanning the same numbers every week, so anomalies jump out faster than they would from a quarterly check-in.

Flow monitoring. At minimum, check your email and SMS flow sends weekly. A flow that stops sending is invisible until you check, and it can cost weeks of revenue before anyone notices.

Spend anomaly alerts. Set up billing and budget alerts in your ad accounts. A card decline or a budget cap hit is one of the most common causes of sudden revenue drops, and it's one of the easiest to prevent.

A simple change log. Keep a running note of what changed each week: ads launched, flows edited, code deployed, prices changed. When something breaks, the change log is the first thing you'll check. Without it, you're guessing.

A revenue baseline. Know what normal looks like. What's a normal Monday, a normal Thursday, a normal weekly total? Founders who have a felt sense of "normal" spot abnormal faster.

This is the gap Arlo was built to fill. We pull your Shopify data every week and flag the anomalies that would take a manual review to catch. The weekly report surfaces what changed, what's trending, and where the biggest preventable leak is, so you're running the diagnostic framework above on a schedule rather than when something is already on fire. You still do the investigation, we just make sure you know to start.

The framework works with or without the tool. What matters is that someone in the business is actually watching.

#Frequently Asked Questions

How big does a weekly drop need to be before I investigate?

For businesses above $50K per week, anything over 15% week-over-week is worth a 30-minute check. Below that, your normal volatility is probably wider and you need a bigger trigger (maybe 25%). The goal is to catch real problems, not to chase noise.

What if I can't find a change event?

Widen the window. Some drops are delayed. A flow paused three weeks ago doesn't bite until week three. If you've checked 14 days and found nothing, check 30. If you've checked 30 and found nothing, the cause is probably external (seasonal, category, competitor) and the question shifts from "what changed" to "how long will this last."

Should I pause spend during a drop?

Only if you can specifically identify why. Pausing working campaigns because revenue is down (without knowing it's a paid problem) often makes the drop worse. If diagnosis shows an acquisition issue and the specific campaign is clearly underperforming, pause it. Otherwise, hold.

How do I know if it's a market-wide drop or a me-problem?

Check two or three peers if you have access to them (founder groups, shared data, public earnings). Check Google Trends for your category. If the drop shows up everywhere, it's category-wide and your response should be conservative. If only you're seeing it, it's a you-problem and you need to find the change event.

What if both acquisition and retention are down?

Usually one of two things. Either the site itself has a problem (checkout broken, pixel down, inventory gap) affecting everyone, or there are two unrelated problems happening at once. Start with site-level checks first, then split the two problems and investigate each separately.

How long should the full investigation take?

The 30-minute first pass should produce a hypothesis. The 48-hour deeper investigation should confirm it. If you're still investigating after three days without a clear answer, you're probably looking at the wrong layer and should go back to Step 1.

What if the drop corrects itself before I finish investigating?

Still investigate. A drop that corrects on its own usually had a cause you'd want to know about for next time. Something went out of stock and came back in. An ad account glitch resolved. A flow error auto-recovered. Any of these can recur, and the second time it happens, you want to recognise the pattern in two minutes instead of two hours.

Do I need to involve my team?

Not for the investigation. For the fix, yes. The diagnostic framework is a founder or senior operator exercise because it requires full context on recent changes, current strategy, and business priorities. Once you have a hypothesis, bring in whoever owns the specific fix (agency, designer, developer, email manager).

If you want your Shopify data monitored weekly and the diagnostic questions answered before the drop compounds, that's what Arlo does. The weekly report points at what changed, what's leaking, and what to investigate first. You still run the framework. We just make sure you catch the drop in week one, not week four.