There is a specific way D2C accounts get destroyed, and it is not by neglect. It is by attention. A marketer opens Ads Manager every morning, sees yesterday's ROAS down 18 percent, and adjusts budgets. The next morning it is up 12 percent, so they adjust again. Within a fortnight every ad set is in permanent learning, performance is genuinely worse, and the marketer concludes the account needs restructuring.
What actually happened is that they were reading noise as signal. In an account doing forty purchases a day, a single day's ROAS swing of 20 percent is entirely ordinary variance — it would occur regularly even if nothing changed at all. Reacting to it is not diligence; it is adding random inputs to a system that was working.
So the useful framing is this: daily tracking and daily decision-making are different activities. You should look at these ten metrics every morning, because the point of daily observation is to catch things that are genuinely broken and to notice trends forming. You should act same-day on very few of them.
Executive Performance Asset
Download Deeptanshu Sharma's Multi-Touch GTM Attribution & Server-Side CAPI Playbook
Get immediate access to pre-built GTM server containers, first-party cookie extenders, and value attribution matrix sheets built for Series A to E companies.
This guide covers the ten metrics that belong on a D2C morning check, what each is actually telling you, the specific thresholds that warrant action, and — the part most metric lists omit — which ones you are allowed to touch today and which you must leave alone.
The ten, and the four you may act on
Watch daily: spend pacing, MER, cost per purchase, new customer share, add-to-cart rate, ATC-to-purchase rate, CPM, frequency, creative spend concentration, and COD confirmation or RTO rate. Act same-day only on: spend pacing anomalies (a runaway or stalled campaign), a collapsed funnel rate signalling a broken site or checkout, a disapproved core creative, and launching new creative. Everything else — budgets, bids, audiences, structure — moves on a three-to-seven-day window with enough conversions behind it to mean something. The discipline of looking daily and acting weekly is what separates accounts that compound from accounts that churn.
Tired of Rising CAC & Attribution Leakage?
Work directly with Deeptanshu Sharma to audit your media strategy, funnel bottlenecks, and server-side tracking.
1. The Efficiency Block (Metrics 1–4)
These four answer the only question that ultimately matters: is the money working. Read them together — individually each can mislead badly.
1. Spend pacing against plan
Yesterday's spend against your daily target, and month-to-date against the monthly budget. Unglamorous and first on the list because it is the only metric where a same-day intervention is almost always correct.
Act when: spend is more than roughly 25 percent above or below plan. Overspend usually means a budget edit or a campaign exiting learning aggressively; underspend usually means low bids, a small audience, or a disapproved creative starving the ad set. Both are real and both are fixable today.
2. MER — blended marketing efficiency ratio
Total revenue from your own backend divided by total ad spend across every platform. The single most important number on this list, because it is the only efficiency figure that cannot be inflated by attribution. Meta will claim revenue, Google will claim the same revenue, and the sum will exceed what your bank received. MER never does.
Act when: the seven-day rolling MER breaks below your contribution-margin break-even. Not on a single day — a one-day MER movement is almost pure noise. Know your break-even number before you need it.
3. Cost per purchase, and platform ROAS beside it
The in-platform view. Useful for comparing campaigns against each other, unreliable as an absolute measure. Track the gap between Meta-reported purchases and actual backend orders daily — if that gap changes suddenly, your tracking broke, not your marketing.
Act when: the Meta-to-backend order gap shifts materially in a day. That is a tracking incident and it is urgent, because every optimisation decision downstream is now running on corrupted data.
4. New customer share and nCAC
What proportion of yesterday's orders came from first-time buyers, and what each cost. This is where D2C accounts quietly rot: ROAS holds steady while the mix shifts toward existing customers who would have bought anyway. Acquisition performance is deteriorating and the headline number conceals it perfectly.
Act when: new customer share trends down over two weeks while ROAS is flat. Check retargeting budget share first — it has usually crept up, because retargeting always looks efficient and is harvesting demand other campaigns created.
2. The Funnel Block (Metrics 5–6)
These two localise a problem. When efficiency drops, the funnel rates tell you whether the cause is the traffic you bought or the experience you sent it to — which are entirely different fixes.
5. Add-to-cart rate
Share of product page visitors who add to cart. Primarily a signal about traffic quality and product-page persuasion. If it falls while CTR holds, you are attracting clicks from people the product does not suit — usually a creative or audience mismatch rather than a site problem.
Act when: it drops sharply overnight. That is a site incident — a broken variant selector, an out-of-stock cascade, a price display error — and it is one of the few genuine same-day emergencies on this list.
6. Add-to-cart to purchase rate
Of people who added to cart, how many bought. This isolates checkout from everything upstream. A stable ATC rate with a collapsing ATC-to-purchase rate is almost never a media problem — it is shipping costs revealed late, a payment gateway failing, a coupon field not working, or a checkout step erroring on one browser.
Act when: it falls more than roughly 20 percent day over day. Test your own checkout on a phone before touching a single campaign setting.
The diagnostic these two enable
When ROAS drops, read the funnel rates before deciding anything. CTR down, ATC rate steady means creative fatigue — the ad stopped earning attention. CTR steady, ATC rate down means audience or product mismatch — you are buying the wrong people. ATC steady, purchase rate down means the site or checkout broke, and no campaign change will help. Three identical-looking ROAS declines, three completely different responses. Most accounts skip this and adjust budgets for all three.
3. The Delivery Block (Metrics 7–9)
These three are your leading indicators. They move days before revenue metrics do, which makes them the most useful things on a daily dashboard and the most commonly ignored.
7. CPM
The cost of reaching a thousand people. Rising CPM means one of three things: auction competition increased, your audience is saturating, or Meta rates your creative as less relevant. All three predict a ROAS decline that has not arrived yet, which is precisely why this belongs on a daily view.
Act when: a sustained rise over five to seven days. Seasonal spikes around sale periods and festivals are normal and will revert — do not restructure an account because of an auction event everyone is experiencing.
8. Outbound CTR and hook rate
Use outbound CTR rather than the all-inclusive figure, which counts expands and reactions. For video, watch hook rate — the share still watching at three seconds — because it isolates whether the opening earns attention, which is the one thing you can actually fix in the next creative.
Act when: a steady decline across several days on a previously strong creative. That is fatigue, and the response is new creative, not a bid change.
9. Frequency and creative spend concentration
Frequency tells you how exhausted the audience is. Concentration tells you how exposed you are: if 80 percent of spend runs through one creative, you have a single point of failure with no replacement tested. When it fatigues, performance falls off a cliff and you need two weeks to recover.
Act when: concentration exceeds roughly 70 percent on one asset. The action is not to pause the winner — it is to get replacements into testing this week, while it is still working.
4. Metric 10: COD Confirmation and RTO Rate
If cash on delivery is a meaningful share of your orders, this is arguably the second most important metric on the entire list, and it exists nowhere in Meta's reporting.
The problem is straightforward. Meta records a purchase at order placement and attaches full order value. Days later the courier attempts delivery and the customer refuses. You have paid forward shipping, you will pay reverse shipping, you hold returned inventory, and you received nothing — while Meta has spent the intervening week optimising toward more people who resemble that customer.
Track it by campaign and audience, not just in aggregate
- Confirmation rate — share of COD orders verified by OTP or confirmation call. Available within a day, so it is genuinely a daily metric.
- RTO rate by campaign — lags by the delivery cycle but must be attributed back to the campaign that generated the order, or you cannot act on it.
- Prepaid share — the proportion choosing to pay upfront, which is a strong proxy for intent and varies noticeably by creative and audience.
- Effective ROAS — platform ROAS adjusted for your realised delivery rate. This number frequently reorders which campaigns are actually profitable.
Refusal rates vary sharply between campaigns, audiences and creatives — heavy discounting and urgency-led creative tend to produce more impulsive orders and more refusals. A campaign with the best ROAS in the account and a poor delivery rate can be your least profitable, and you will never discover it from Meta's interface. Feeding delivery outcomes back through the Conversions API is how you stop the algorithm optimising toward customers who never pay; our D2C CAPI guide covers the implementation.
5. The Fifteen-Minute Morning Ritual
A repeatable sequence beats an unstructured scroll through Ads Manager, mostly because it stops you finding a reason to change something every day.
| Minutes | Check | Only if |
|---|---|---|
| 0–2 | Anything broken? Spend anomalies, disapprovals, zero-delivery ad sets | Fix immediately — these are real emergencies |
| 2–5 | Backend orders vs Meta-reported purchases | Gap changed? Tracking incident — escalate today |
| 5–8 | 7-day rolling MER and new customer share | Below break-even for 3+ days? Plan a change, do not make it now |
| 8–11 | Funnel rates: ATC and ATC-to-purchase | Collapsed? Test the site yourself before touching campaigns |
| 11–13 | CPM, frequency, hook rate trends over 7 days | Deteriorating? Brief new creative — not a bid change |
| 13–15 | Creative concentration and COD confirmation rate | Concentration high? Get replacements into test this week |
Notice how many rows end in "plan a change" rather than "make a change." That is deliberate, and it is the hardest part of this discipline to sustain — particularly when someone senior is watching yesterday's number and expects visible activity in response to it.
6. What You May Change Daily, and What You May Not
| Action | Cadence | Why |
|---|---|---|
| Pause a broken or disapproved asset | Immediately | Not an optimisation, a repair |
| Launch new creative | Daily is fine | Adding tests does not disturb existing learning |
| Budget changes | Every 3–7 days, modest steps | Large or frequent edits re-trigger learning |
| Pausing an underperforming ad set | After 50+ conversions or 7 days | Below that you are reading noise |
| Audience or structural changes | Weekly at most | Resets learning entirely; needs a real hypothesis |
| Rebuilding the account | Almost never | Usually recreates the same setup with new IDs and lost history |
The noise test, before any decision
Before acting on any number, ask how many conversions sit behind it. An ad set with nine purchases yesterday tells you almost nothing — the confidence interval around that figure is enormous, and a 30 percent swing is entirely expected variance. If you would not bet your own money on the difference being real, do not spend the company's on it. This one question prevents most of the damage marketers do to their own accounts.
7. Building the View: Where Each Number Comes From
The reason most D2C teams do not run this dashboard is not disagreement with the metrics — it is that four of the ten do not exist in Ads Manager, so assembling the view requires pulling from three places. Worth doing once properly rather than reconstructing manually every morning.
| Metric | Source | Assembly note |
|---|---|---|
| Spend, CPM, CTR, frequency | Meta | Native; use outbound CTR, not the inclusive figure |
| Funnel rates | GA4 or your store platform | Store platform is more reliable; GA4 lets you split by source |
| MER and actual revenue | Store backend | Never platform-reported. This is the whole point of the metric |
| New vs returning customer | Store backend | Define "new" once — first order ever, not first this month |
| COD confirmation and RTO | Order management or logistics | Must be attributed back to the originating campaign to be actionable |
| Creative concentration | Meta, aggregated | Top asset spend as a share of total; needs consistent naming |
Two assembly decisions cause most of the trouble later. First, define "new customer" once and write it down — first order ever versus first order in the period produces materially different numbers, and teams routinely discover mid-quarter that two reports used different definitions. Second, attribute RTO back to the campaign rather than reporting it as a company-wide rate. An aggregate refusal rate is a logistics statistic; a per-campaign one is a marketing decision.
On tooling, resist over-engineering this. A connector feeding Looker Studio covers it for most brands, and a spreadsheet updated by a scheduled export covers it for smaller ones. The value is in looking at the ten numbers together every morning, not in the sophistication of what renders them — and a dashboard project that takes six weeks has cost you six weeks of not looking.
Automate the four that are genuinely urgent
A dashboard requires someone to open it, and mornings get missed. The four same-day emergencies deserve alerts rather than relying on the ritual, because each of them costs money for every hour it runs unnoticed:
- Spend anomaly. Daily spend more than roughly 25 percent above or below plan. Catches runaway campaigns and starved ad sets equally.
- Order volume anomaly. Backend orders more than 30 percent below the trailing average for that weekday. This is the single most valuable alert on the list because it catches broken checkouts, failed payment gateways and tracking outages simultaneously.
- Meta-to-backend divergence. The ratio between platform-reported purchases and actual orders shifting materially. Almost always a tracking incident, and everything downstream is corrupted until it is fixed.
- Disapproval on a top-spending creative. Concentration means one disapproval can starve an ad set within hours.
Note that three of those four alerts are anomaly detections rather than threshold breaches — they compare against a trailing baseline rather than a fixed number. That distinction matters in D2C, where weekday patterns and promotional periods make fixed thresholds fire constantly and get muted within a fortnight. An alert that everyone has learned to ignore is worse than no alert, because it creates the impression of coverage that does not exist.
8. Pros and Cons of Daily Monitoring
| Pros | Cons |
|---|---|
| Broken tracking and broken checkouts get caught in hours, not weeks. | Daily visibility creates constant pressure to intervene. |
| Leading indicators give a week of warning before revenue drops. | Leading indicators are noisy and easy to over-read. |
| Runaway spend is contained same-day rather than at month end. | Fifteen minutes daily is real time that adds up. |
| Creative fatigue is spotted while there is time to replace it. | Encourages replacing creative that was still working. |
| MER keeps everyone anchored to real revenue, not claimed revenue. | Blended metrics hide which specific channel caused a change. |
| COD and RTO tracking reveals unprofitable winners. | Requires backend data Meta will never provide. |
9. Advantages and Disadvantages in Practice
What a disciplined daily habit produces
- Incidents stop becoming disasters. A checkout broken on Friday evening found on Saturday morning costs one day of spend rather than a weekend of it.
- Creative production gets planned, not panicked. Watching concentration and hook rate means replacements are briefed while the winner still works, rather than after it collapses.
- Arguments about performance shorten. When MER and new-customer share are on the same daily view, "Meta is working" becomes a checkable claim rather than a debate about attribution.
- Fewer changes produce better results. Accounts run on a look-daily, act-weekly rhythm consistently outperform accounts edited every morning, because they stay out of permanent learning.
Where daily tracking backfires
- Reacting to variance. The single most damaging habit in performance marketing. A bad Tuesday is usually just a Tuesday.
- Leadership reading the daily view. A founder watching yesterday's ROAS will ask for action on noise. Give them the seven-day rolling view instead, and explain why once, properly.
- Optimising to platform-reported numbers. Meta's ROAS is not your ROAS. Anchoring to it produces confident decisions on inflated figures.
- Killing creative during learning. New creative frequently underperforms for its first days by design. Judge on conversion volume, not on elapsed time.
- Weekday and seasonal effects misread as trends. Most D2C accounts have strong day-of-week patterns. Compare Tuesday to Tuesday, not Tuesday to Sunday.
10. Myths and Facts
| Myth | Fact |
|---|---|
| Checking daily means optimising daily. | Daily observation catches breakage. Daily editing keeps ad sets in learning and makes performance worse. |
| Platform ROAS is the headline metric. | Every platform over-claims the same conversions. MER, from your own revenue, is the only figure that reconciles with the bank. |
| A ROAS drop means the campaign stopped working. | It might be creative fatigue, a site failure, an auction event or ordinary variance. The funnel rates tell you which. |
| Rising CPM is out of your control, so ignore it. | It is your earliest warning of saturation and creative decay, and it moves days before revenue does. |
| Stable ROAS means acquisition is healthy. | Not if new customer share is falling. Retargeting can hold ROAS steady while genuine acquisition quietly deteriorates. |
| Orders placed is the number that counts. | In COD markets, orders delivered is. A campaign with great ROAS and poor delivery can be your least profitable. |
| One winning creative means you are in good shape. | It means you have a single point of failure. Concentration above roughly 70 percent is a risk to manage, not a success. |
| If yesterday was bad, do something today. | Ask how many conversions sit behind the number first. Below about fifty, you are reading noise and acting on it will cost you. |
Look at all ten every morning and act on almost none of them. Spend pacing, a changed gap between Meta-reported and backend orders, a collapsed funnel rate and a disapproved creative are genuine same-day emergencies. Everything else — budgets, bids, audiences, structure — waits for a window with enough conversions behind it to mean something, which in most D2C accounts is three to seven days. Anchor to MER rather than platform ROAS, because Meta and Google will both claim the same revenue and only your backend knows the truth. Watch new customer share, because that is where acquisition quietly rots while ROAS looks fine. Track delivery rate if you sell cash on delivery, because orders placed is not revenue. And before every decision, ask how many conversions sit behind the number — the discipline of not acting on a bad Tuesday is worth more than any optimisation you will make this quarter.