---
title: "Stop Reading Meta Agency Case Studies. Ask for the Change Log."
description: "ROAS screenshots show an agency’s best moment, not how it will run your account. Judge its first month by the changes it makes, the reasons behind them, and what your business actually banks."
url: "https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/2d629e42-f0a4-4e9a-bff2-9fff9bf86ab9.png"
published: "2026-11-01T06:04:27.830Z"
modified: "2026-11-01T06:04:27.900Z"
---

November 1, 2026 · 9 min read

# Stop Reading Meta Agency Case Studies. Ask for the Change Log.

[Alexander PerelmanHead Of Product @ groas](https://groas.com/author/alexander-perelman)[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![A shiny gold "5x" trophy revealed as a hollow cardboard prop stands beside a worn, coffee-stained notebook full of dated handwritten changes.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/2d629e42-f0a4-4e9a-bff2-9fff9bf86ab9.png)

In this article

1. [A case study is a trophy, not an operating record](#a-case-study-is-a-trophy-not-an-operating-record)
2. [Buy the decisions you can inspect](#buy-the-decisions-you-can-inspect)
3. [ROAS is an easy target to hit the wrong way](#roas-is-an-easy-target-to-hit-the-wrong-way)
4. [Make month one the interview](#make-month-one-the-interview)
5. [Checklist: the first-month log worth paying for](#checklist-the-first-month-log-worth-paying-for)
6. [Who should skip the log test](#who-should-skip-the-log-test)

Stop reading Meta agency case studies. The only evidence that matters when you hire one is what it changes in your account next Tuesday.

A good month can make a beautiful PDF. Black Friday does the heavy lifting, the agency picks the right 14 days, and a tight crop turns a temporary spike into a 5x ROAS trophy. Now say you spend $3k a month selling an $80 product with real margins to protect. That trophy came from a different product, budget, margin, and attribution setting. It tells you what happened once for someone else. **It does not tell you what the agency will do with your money.** For that, ask for its change log and make the first month the interview.

## A case study is a trophy, not an operating record

Most Meta case studies follow a familiar pattern. An agency picks a short-window spike, often Black Friday or a clearance sale where past buyers flood back, lets Meta take credit for the purchases, then packages the result as proof it can scale an account year-round. One breakdown of fashion ecommerce case studies describes [curated spikes presented as repeatable growth](https://www.elitebrands.org/blog/agencies-meta-case-studies-fashion-ecommerce). The sale created demand. The ads stood next to the register.

That is how the ROAS headline can be true and useless at the same time. The same breakdown describes a $4M ARR streetwear brand shown at [6.5x ROAS on winter clearance while its Shopify backend showed a negative 4% net margin](https://www.elitebrands.org/blog/agencies-meta-case-studies-fashion-ecommerce). The campaign spent $15k buying heavily discounted repeat customers. If you sell at full price and need new customers, that example does not answer your hiring question.

I used to tell clients to look for high ROAS proof. I was wrong. Without the margin, the offer, the buyer mix, and the attribution window, the number cannot tell you whether the agency knows how to run your account in an ordinary February. Ask for those details and the trophy starts looking less like evidence and more like décor.

![A framed 5x ROAS screenshot gathering dust beside an open weekly work log](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/51ac87e9-bc67-4537-b0eb-f0fdecc87a86.png)

### A 5x headline hides three decisions

First, **the attribution window changes what Meta gets to claim**. Under a 7-day click plus 1-day view window, a purchase can receive credit after a click up to a week earlier or a view the day before. That credit is not the same thing as proving the ad caused the purchase. One discussion of platform reporting puts the gap between [credited revenue and banked revenue at 20–40%](https://www.adeptecom.com/blog-meta-ads-roas-truth.html). Change the reporting window to 1-day click, and displayed ROAS can appear to [fall 30–50% without a change in spend](https://jetfuel.agency/meta-ads-attribution-settings-in-2026-windows-best-practices-and-what-dtc-brands-get-wrong/). The business did not suddenly lose those sales. The reporting rule changed.

Second, the comparison period can move under your feet. As of January 12, 2026, Meta stopped returning 7-day and 28-day view data in the Insights API; in March 2026, click credit narrowed to link clicks, with likes and comments moved aside. One audit illustrates how far apart reported figures can sit: [Meta showed 3.23x against 0.93x on a click-only measure for the same spend](https://www.adbeacon.com/meta-attribution-bfcm-2026/). If a case study combines periods or omits its window, you cannot tell what changed in the account and what changed in the ruler.

Third, a headline needs a baseline. Triple Whale full-year 2025 data covering about 35,000 ecommerce brands puts [median Meta ROAS at 1.93x](https://eightx.co/blog/average-ecommerce-roas-by-vertical-2026-meta-google-benchmarks). A 5x example is an outlier, which is precisely why it made the agency website. It may be a real result. It is still a poor forecast for a different account.

Before you admire the number, ask what window produced it, what the store or CRM recorded, and what margin remained. If the agency cannot answer, stop reading the slide.

## Buy the decisions you can inspect

I judge media buying the way I judge ad copy: by the work around the winner, not just the winner. A case study shows you an outcome after the agency has chosen the account, period, and metric. A change log shows you the process you are about to pay for: what it tested, what it killed, what it moved, and why. **The log is evidence of management. The screenshot is evidence that a screenshot exists.**

Stagnant management has a recognizable shape. One audit of neglected accounts describes [an always-on sales campaign that ran for more than a year with only 16 ads](https://ncmborz.com/blog/signs-ads-agency-wasting-budget.html). After a takeover, the new team ran 44 campaigns versus the prior four over eight months. In a smaller-budget account in the same write-up, an active team [tested 118 creatives in 10 months on $5 to $15 a day](https://ncmborz.com/blog/signs-ads-agency-wasting-budget.html). Those counts are not targets for every account. They show the difference between a team making decisions and one collecting a fee while yesterday’s ads keep running.

That is why I look at creative velocity before I trust a ROAS claim. Motion 2026 benchmarks found brands testing [10 or more concepts a month had 31% lower CPA than brands testing fewer than five](https://scigrowth.com/blog/meta-ads-creative-testing-framework-for-dtc-brands/). The working rule cited there is 20 to 25 concepts to find three winners because only 10 to 20% scale. A log lets you see whether the agency has a way to find winners, or whether its entire plan is to keep asking the last winner to work harder.

Do not turn those counts into another trophy. Ask how many new concepts launched last month, what each was meant to learn, and which ones the agency stopped spending on. A list of launches without decisions attached is an activity report. A list of decisions with reasons is management.

## ROAS is an easy target to hit the wrong way

Searching for a Meta ads management company that “focuses on ROAS” sounds sensible. It can select for the team most willing to make platform ROAS look good while your business stands still. If a buyer answers only to credited sales, retargeting past buyers is attractive. So is a generous attribution window. So is leaving warm traffic inside Advantage+ and presenting the resulting number as proof of new demand.

A buyer who answers to banked revenue has harder questions to face. Did the store actually take in more? Which sales came from new customers? What happened to acquisition cost? Did tracking count the same purchase twice? The measurement approach other operators recommend includes [blended ROAS or MER, CAC by cohort against LTV, Pixel and Conversions API deduplication, and holdouts to assess incremental lift](https://www.adeptecom.com/blog-meta-ads-roas-truth.html). Those checks cannot make an unprofitable campaign profitable. They can stop a flattering dashboard from concealing one.

I learned enough from late-night search-term mining to distrust work that comes wrapped in jargon. What a deck calls “synergy,” I call two people doing the same job. What a deck calls “performance,” I want to see in the store or CRM. If you spend $2k to $5k a month, you do not need the agency with the highest screenshot. You need the one willing to put store revenue beside total spend, state the attribution window in writing, and keep it consistent while it makes decisions.

**If they will not name the window, they are selling you the window.**

## Make month one the interview

Stop interviewing the deck. Interview the work. Any team can talk for 45 minutes about its best campaign. Ask it to show four straight weeks of changes, including the ideas that failed, and the conversation gets useful quickly. You can see whether anyone noticed a tired creative, changed a budget for a stated reason, checked the result, and decided what to do next. You can also see the weeks when nothing happened.

That is why the contract matters more than the case study. A six-month commitment pays the agency whether it tests or coasts. A month-to-month arrangement makes the next invoice depend on work you can inspect. It does not guarantee a profitable first month. It gives you a way to distinguish an agency learning your account from one waiting out the contract. **Exit power makes a quiet account harder to excuse.**

![A weekly Meta ad change log beside a long-term contract](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/ed9a37e4-efa3-43d1-a2f3-4584f1910975.png)

This is where groas fits my bias. Its [Meta Ads service costs $1,499 a month, flat and month to month, with no cut of spend or setup fee](https://groas.com/pricing). It says the service goes live in your account within 24 hours of onboarding, with audiences, budgets, and ads worked around the clock. I like that structure because month one is checkable: you can look for ads tested, winners scaled, and decisions explained, with a dedicated manager who has to earn month two. Do not take the structure as a substitute for looking. Use it to make looking possible.

What weekly reporting should show is boring on purpose: date, change, reason, result, next check. Not a stack of impressions. Not a ROAS screenshot with the window cropped away. Open change history and count the new ads in the last 90 days. Then put Ads Manager revenue next to what Shopify or your CRM recorded and ask the agency to explain the gap in writing. Those figures need not match; the explanation needs to exist. Otherwise, you may be paying for sales the platform credited rather than sales the work created.

Never sign a term that prevents you from firing a quiet account.

## Checklist: the first-month log worth paying for

Take this to the sales call, then use it again after week four. Ask for a recent log from a business like yours and insist on the same record for your account. If the agency cannot show these decisions, you are buying its trophy cabinet.

1. **Creative tests launched and retired, with kill reasons.** Record the concept, launch date, spend, result, and decision. One new ad a month is not a testing process.
2. Tracking verified in week one. Record [Pixel and Conversions API deduplication checks and the attribution window in writing](https://www.adeptecom.com/blog-meta-ads-roas-truth.html); reconcile reported revenue with what the business banked.
3. Budget and audience moves with reasons. Record what shifted between prospecting and retargeting, what triggered the move, and when it will be checked again.
4. **What the agency chose not to change, and why.** Record the reason for leaving something alone and the date it will be revisited. Skip this line and every quiet week can masquerade as careful management while another retainer comes due.

## Who should skip the log test

This advice has a narrow target. If you only need someone to make UGC and product videos while your team runs media, judge the portfolio. The reel is the work. If you spend well into six figures a month across five countries with legal review on every claim, you need procurement, security review, and a custom scope, not just a first-month log. I am writing for the owner spending $2k to $10k a month who needs someone to run Facebook and Instagram ads. For that buyer, the trophy cabinet is a distraction.

Keep choosing by case study and here is what happens. You sign the six-month term because the 5x slide felt safe. By month two, the account is quiet: the same four campaigns, the same tired winners, and dashboard ROAS holding up while retargeting takes credit for repeat buyers. Shopify tells a different story, but nobody puts the numbers side by side. By month four, you are paying a retainer for change history with nothing in it. The case study did not warn you because it was never a record of the work you were buying.

Send this before your next sales call: “Show me last month’s change log for an account spending within $2k of my budget, with the attribution window printed at the top.” Ask what was launched, killed, left alone, and revisited. If they send screenshots instead, you have your answer. The agencies worth paying can explain their decisions. Everyone else needs you to keep staring at the trophy.

## Frequently Asked Questions

### Why should I be skeptical of high ROAS numbers shown in Meta agency case studies?

High ROAS case studies often reflect short-term spikes, like Black Friday sales, rather than consistent year-round performance. These figures typically lack crucial context such as margin, offer, buyer mix, and attribution settings, which prevents them from reliably predicting success for a new account.

### How does changing the attribution window impact reported ROAS?

Changing the reporting window can make displayed ROAS fall significantly without any change in ad spend. Moving from a 7-day click plus 1-day view window to a 1-day click window can cause reported ROAS to fall by 30 to 50%. This is a change in the reporting rule, not a loss of sales.

### What should I consider when evaluating the baseline of an agency's case study results?

A high ROAS headline needs a baseline, and an exceptional example is often a statistical outlier. Data covering 35,000 e-commerce brands showed the median Meta ROAS is 1.93x. Evaluating a single high number without this context makes it an inadequate forecast for your business.

### What is the functional difference between viewing an agency's change log and reviewing a case study?

A case study presents a final outcome after the agency has selected the account, period, and metric for presentation. A change log provides evidence of management by detailing what the agency tested, what it killed, what it moved, and the specific reasons behind those decisions.

### How do I determine if an agency is actively managing an account versus simply collecting a fee?

Agencies that test a high volume of concepts a month tend to perform better. Benchmarks found that brands testing 10 or more concepts a month had 31% lower CPA than those testing fewer than five. A change log allows you to verify if the agency possesses a way to find winning ads.

### What measurement methods should a business use to assess true campaign profitability?

Businesses concerned with bank revenue should use measurements like blended ROAS, CAC by cohort against LTV, and utilizing holdouts to assess incremental lift. These necessary checks prevent a flattering dashboard from concealing an unprofitable campaign.

## Paid Ads & SEO, Run For You

Businesses pay one flat monthly fee per service, agencies resell it under their own brand, and groas does the work either way.

[See Pricing](https://groas.com/pricing)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","description":"groas is a done-for-you marketing service, not software: elite people working with proprietary technology systems run search ads, Meta ads, SEO and AI search for businesses, and white label for agencies, with a dedicated senior strategist owning every account.","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#webpage","url":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s","name":"Stop Reading Meta Agency Case Studies. Ask for the Change Log.","description":"ROAS screenshots show an agency’s best moment, not how it will run your account. Judge its first month by the changes it makes, the reasons behind them, and what your business actually banks.","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/2d629e42-f0a4-4e9a-bff2-9fff9bf86ab9.png"},"breadcrumb":{"@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#breadcrumb"},"dateModified":"2026-11-01T06:04:27.900Z","mainEntity":{"@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"Stop Reading Meta Agency Case Studies. Ask for the Change Log.","item":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#article","url":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s","mainEntityOfPage":{"@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"Stop Reading Meta Agency Case Studies. Ask for the Change Log.","description":"ROAS screenshots show an agency’s best moment, not how it will run your account. Judge its first month by the changes it makes, the reasons behind them, and what your business actually banks.","datePublished":"2026-11-01T06:04:27.830Z","dateModified":"2026-11-01T06:04:27.900Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/2d629e42-f0a4-4e9a-bff2-9fff9bf86ab9.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"agency, month, account, meta, case, roas, change, log","wordCount":1959,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"My background includes Goldman Sachs and Stanford Computer Science.","image":"https://groas.com/media/blog/4f17cac3a81acc224cc1b2cfab8f94bbc295a3089d13e8421125771ef8d4064f.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"},"sameAs":["https://www.linkedin.com/in/alexander-433793253/"]},{"@type":"FAQPage","@id":"https://groas.com/post/stop-reading-meta-agency-case-studies-ask-for-last-month-s#faq","mainEntity":[{"@type":"Question","name":"Why should I be skeptical of high ROAS numbers shown in Meta agency case studies?","acceptedAnswer":{"@type":"Answer","text":"High ROAS case studies often reflect short-term spikes, like Black Friday sales, rather than consistent year-round performance. These figures typically lack crucial context such as margin, offer, buyer mix, and attribution settings, which prevents them from reliably predicting success for a new account."}},{"@type":"Question","name":"How does changing the attribution window impact reported ROAS?","acceptedAnswer":{"@type":"Answer","text":"Changing the reporting window can make displayed ROAS fall significantly without any change in ad spend. Moving from a 7-day click plus 1-day view window to a 1-day click window can cause reported ROAS to fall by 30 to 50%. This is a change in the reporting rule, not a loss of sales."}},{"@type":"Question","name":"What should I consider when evaluating the baseline of an agency's case study results?","acceptedAnswer":{"@type":"Answer","text":"A high ROAS headline needs a baseline, and an exceptional example is often a statistical outlier. Data covering 35,000 e-commerce brands showed the median Meta ROAS is 1.93x. Evaluating a single high number without this context makes it an inadequate forecast for your business."}},{"@type":"Question","name":"What is the functional difference between viewing an agency's change log and reviewing a case study?","acceptedAnswer":{"@type":"Answer","text":"A case study presents a final outcome after the agency has selected the account, period, and metric for presentation. A change log provides evidence of management by detailing what the agency tested, what it killed, what it moved, and the specific reasons behind those decisions."}},{"@type":"Question","name":"How do I determine if an agency is actively managing an account versus simply collecting a fee?","acceptedAnswer":{"@type":"Answer","text":"Agencies that test a high volume of concepts a month tend to perform better. Benchmarks found that brands testing 10 or more concepts a month had 31% lower CPA than those testing fewer than five. A change log allows you to verify if the agency possesses a way to find winning ads."}},{"@type":"Question","name":"What measurement methods should a business use to assess true campaign profitability?","acceptedAnswer":{"@type":"Answer","text":"Businesses concerned with bank revenue should use measurements like blended ROAS, CAC by cohort against LTV, and utilizing holdouts to assess incremental lift. These necessary checks prevent a flattering dashboard from concealing an unprofitable campaign."}}]}]}
```