---
title: "Why Meta ROAS Dropped: Diagnose It in Just 6 AI Prompts"
description: "Six ordered prompts for your AI connected through groas to check Meta tracking, auction costs, creative, audiences, offers and landing pages before changing ads."
url: "https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/d7930154-845f-4dc9-825f-79682dbd6c32.png"
published: "2026-11-19T13:06:44.472Z"
modified: "2026-11-19T13:06:44.527Z"
---

November 19, 2026 · 10 min read

# Why Meta ROAS Dropped: Diagnose It in Just 6 AI Prompts

[](https://groas.com/author/alexander-perelman)

[Alexander Perelman](https://groas.com/author/alexander-perelman)Personally Tested

Head Of Product @ groas

[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![A clay marketer frantically repaints a shop-window ad while the unplugged coin counter behind him lets sales pile up uncounted on the floor.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/d7930154-845f-4dc9-825f-79682dbd6c32.png)

In this article

1. [Before the six prompts: connect the evidence](#before-the-six-prompts-connect-the-evidence)
2. [1. Check whether the ROAS drop is real](#1-check-whether-the-roas-drop-is-real)
3. [2. Separate auction inflation from weaker ads](#2-separate-auction-inflation-from-weaker-ads)
4. [3. Find fatigue at the ad level](#3-find-fatigue-at-the-ad-level)
5. [4. Check where the budget moved](#4-check-where-the-budget-moved)
6. [5. Inspect the offer and the page after the click](#5-inspect-the-offer-and-the-page-after-the-click)
7. [6. Approve one fix, not an account renovation](#6-approve-one-fix-not-an-account-renovation)

You get six prompts to diagnose a Meta ROAS drop, run in order against your real account data. Spend them rewriting ads before checking tracking, and you can make six confident moves in the wrong direction.

I learned this the slow way. I used to open a tanking account and start with creative because creative is visible and changing it feels like work. I burned days rewriting ads that were never the problem.

ROAS is purchase value Meta counted divided by spend, with modeling filling gaps when signal is partial. The suspects are a short list: tracking, auction costs, creative, audience, offer and landing page. This script makes your connected AI check them in order, then propose one fix. No brainstorming detour. No account-wide renovation because Tuesday looked bad.

## Before the six prompts: connect the evidence

If ChatGPT or Claude cannot see your Meta account, the first prompt returns the generic list you could find on any blog. **No connection, no diagnosis.**

I run this through [groas](https://groas.com/), which connects your ad accounts and website to ChatGPT, Claude, Cursor, Gemini or another agent over MCP or with an API key. It also collects competitor ads and the landing pages and funnels behind them, so those checks happen alongside your account analysis rather than in a separate research expedition.

The price is $19 a month flat for a whole team. Autopilot is off until you switch it on; every change is logged with its reason and can be undone in one step. For this script, leave autopilot off and instruct your AI to make no changes without approval.

If you use ChatGPT, complete [Connect Facebook Ads Manager to ChatGPT Without the CSV Shuffle](https://groas.com/post/how-to-connect-facebook-ads-manager-to-chatgpt-without) first. Connect the store and website data needed for the comparisons, too. If a requested source is unavailable, the AI must flag that gap, not manufacture a verdict from Meta alone.

Four things get cut before the first prompt:

- **Open-ended guessing:** “Why did ROAS drop?” without dates or a checking order.
- **Industry benchmarks:** they cannot tell you whether your account makes money.
- **New ad ideas:** premature until tracking and auction costs are checked.
- **Scaling advice:** more spend is not a diagnosis.

Each numbered step below contains one prompt. Run them in order. If a step exposes a measurement problem, stop and resolve it before spending the remaining prompts. Six is the ceiling, not a quota.

![Detective board mapping six Meta ROAS suspects in order](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/22e1d907-240a-4f87-a62b-a3435a1ba980.png)

## 1. Check whether the ROAS drop is real

**Cost: prompt 1 of 6. Buys: a measurement check before a marketing intervention.**

> Identify the date Meta ROAS began falling. Compare Meta purchases and purchase value with Shopify orders and revenue for the 14 days before and after that date. Did the gap between the sources change when ROAS fell, rather than both sources simply moving together? Compare website sessions in the same windows. Check dataset health, pixel and Conversions API deduplication, and attribution-setting changes where that information is available. Flag anything you cannot inspect. Make no changes in this step or any later step without my explicit approval.

The tell is a change in the gap. Meta down and Shopify flat points toward tracking or attribution. Both down with sessions flat points toward the store converting worse. Both down with sessions down points toward a traffic or delivery problem. Those are directions for investigation, not verdicts.

This is why I start with measurement rather than the ad everyone is bored of looking at. ROAS depends on the purchase value Meta counts. If that count changes, the chart can fall without the business deteriorating in the same way. The [tracking-first diagnosis](https://www.prism3.agency/blog/meta-ads-roas-dropped-diagnose-shopify) matters more than a fresh hook.

Watch the comparison itself, too. Attribution definitions and reporting windows can change what gets credited. Do not compare unlike settings and label the difference a performance decline. A neat before-and-after table is useless if its columns mean different things.

If this step exposes an event, deduplication or attribution problem, stop. Resolve it before proceeding. For the longer checklist, use [The Meta Tracking Pre-Flight: 19 Checks Before You Trust ROAS](https://groas.com/post/the-meta-tracking-pre-flight-19-checks-before-you-believe).

**Do not optimize against a number you have not checked.**

## 2. Separate auction inflation from weaker ads

**Cost: prompt 2 of 6. Buys: a distinction between paying more for attention and getting less from it.**

> Compare CPM, CTR and CVR by campaign for the 14 days before and after the ROAS drop. Show the size and timing of each change. Identify which metric moved first, and compare the same period last year if comparable data is available. Separate auction-cost changes from click-through and conversion changes. Do not recommend new creative yet.

I want the mechanism before the recommendation:

- **CPM rises while CTR and CVR hold:** the account is paying more for similar attention and conversion behavior.
- **CPM holds while CTR falls:** creative or delivery mix deserves the next look.
- **CPM and CTR hold while CVR falls:** the post-click experience and offer become stronger suspects.

Several metrics can move together. The point is to stop treating every ROAS decline as an ad-copy emergency, not to pretend one ratio explains an entire account.

Seasonality belongs here. A Q4 auction can get more expensive while the ad itself keeps doing its job. Comparing those 14-day windows with the same period last year gives your AI context before it blames the creative. If that comparison is unavailable, it should say so.

This step does not authorize a budget change. It establishes whether buying the same attention now costs more. **If CPM is the main mover, a new hook is not automatically the fix.** Continue through the remaining checks before choosing an intervention.

## 3. Find fatigue at the ad level

**Cost: prompt 3 of 6. Buys: a specific retirement candidate instead of an account-wide rewrite.**

> Among the top-spending ads, identify those with rising frequency and falling CTR during the 7 days before the drop. Include CPC and each ad’s 7-day CPA against its trailing 30-day average. Separate prospecting from retargeting. Distinguish evidence consistent with creative fatigue from a CTR decline that could reflect a placement or audience shift. Name candidates for replacement, but do not pause or replace anything.

Fatigue is a pattern, not a single red cell. I look for rising frequency, falling CTR and climbing CPC together. A falling CTR alone does not earn a guilty verdict.

The [fatigue thresholds I use](https://topgrowthmarketing.com/how-to-spot-meta-ad-fatigue) are starting flags:

- **Prospecting:** watch frequency around 2.5 to 3.0 in 7 days; consider action at 3.5 and above.
- **Retargeting:** watch around 4 to 5; consider action at 6 and above.
- **Performance:** flag a CTR decline of 25% or more week over week, alongside rising CPC.
- **Economics:** flag an ad whose 7-day CPA exceeds 1.3 times its trailing 30-day average.

I do not use those flags as an automatic kill switch. An ad can cross a frequency threshold without explaining the account’s ROAS decline. The AI needs to connect the deterioration to the ads taking meaningful spend, then check whether delivery changed underneath them.

I used to treat falling CTR as sufficient evidence of fatigue. It is not. **Find the ad and the pattern before commissioning replacements.** Step 4 checks whether the apparent creative problem is actually a distribution problem.

![Claymation exhausted ad character repeated too often beside a fresh unused ad character](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/250ca118-12ee-4219-85ba-5c3cafbadd03.png)

## 4. Check where the budget moved

**Cost: prompt 4 of 6. Buys: a delivery-mix diagnosis rather than a blanket judgment on the audience.**

> Break ROAS, spend share, CPM, CTR and CVR down by placement and audience segment for the 14 days before and after the drop. Which placements or segments gained spend share while returning less? Check whether Advantage+ delivery shifted toward them. Inspect audience overlap and saturation where available, and distinguish a change in delivery mix from deterioration within the same placement or segment. Flag reporting limits rather than filling them in.

Account averages hide movement. Feed can hold steady while another placement takes more budget and converts worse. The overall chart falls, but “the ads stopped working” is still too crude a diagnosis.

I look for two different patterns:

- **Mix shift:** weaker-returning placements or segments take a larger share of spend.
- **Within-segment decline:** the same placement or audience performs worse than it did before.

They lead to different decisions. The first calls for scrutiny of delivery and placement controls. The second sends me back to the creative evidence or forward to the offer and page. Neither proves the whole audience is exhausted.

Advantage+ can change delivery, but that is not the same as saying it always chases the cheapest click. Check what actually moved in this account. The account data should carry the argument, not a theory about what the algorithm likes.

Overlap matters for the same reason. People exposed through overlapping campaigns can see more advertising than a single campaign’s frequency suggests. **Check exposure across the account before declaring one ad the problem.**

## 5. Inspect the offer and the page after the click

**Cost: prompt 5 of 6. Buys: a check on both your proposition and the alternatives customers see.**

> Use groas competitor data from the last 30 days to compare rival hooks, angles, offers, prices, landing pages and funnels. Identify changes around my ROAS drop date, and distinguish observed changes from inferred performance. Then compare my landing-page sessions, CVR and average order value for the 14 days before and after the drop. Flag page, price, shipping, offer or checkout changes where records are available. Explain whether the evidence points to my offer, my page, a competitor change or an unresolved gap. Do not treat competitor activity as proof of causation.

This is one prompt because the commercial question is one question: did the proposition after the click get worse, either on its own or against what customers could buy elsewhere?

The [Meta Ad Library](https://www.revvyscales.com/blog/facebook-ad-library) shows active ads, copy, landing URLs and start dates. It does not show clicks, conversions, sales or ROAS. Long runtime is not proof of profit. Repeated versions of an angle are a hint worth inspecting, not a competitor’s bank statement.

I use groas here because it collects rival ads on Google and Meta and the landing pages and funnels behind them. The AI can examine that evidence alongside account performance instead of ending with “competition increased,” which is barely a sentence, never mind a diagnosis.

Then inspect your own page. If sessions and CTR hold while CVR falls, check the price, shipping, headline, pop-up and checkout before asking for a new campaign. If purchase value falls, average order value belongs in the explanation too.

A competitor’s changed offer may be relevant. A checkout problem may be decisive. **Fix a broken buying experience before paying to send it more traffic.** New creative aimed at the same broken page just buys more evidence.

## 6. Approve one fix, not an account renovation

**Cost: prompt 6 of 6. Buys: one traceable decision tied to the evidence.**

> Rank the causes supported by steps 1–5. Name the best-supported cause and any unresolved evidence gaps; do not force a single cause if the data supports several or none. Propose one fix, explain its mechanism, and state which CPA or ROAS movement would support the diagnosis. Give an expected effect only if the available evidence supports it. Specify what to monitor and when to review it. Wait for my approval before changing anything. If approved, log the change and its reason so it can be undone.

This is where diagnosis scripts usually turn into renovation projects. The AI rewrites creative, changes targeting and moves budgets. Something improves. Nobody knows why. The deck calls that optimization; I call it losing the control.

**One approved change preserves something you can learn from.** groas logs changes with their reasons and lets you undo them in one step. Use that record. Do not approve three unrelated changes because the explanation sounds confident.

Seven days without a ROAS recovery is a reason to review the evidence, not an automatic instruction to reverse everything. Compare what changed with what the diagnosis predicted. If the expected mechanism did not appear, revisit the cause rather than inventing a success story around a different metric.

Some causes remain outside this script’s reach. It cannot remove iOS signal loss, make seasonal auctions cheap or turn incompatible attribution definitions into a clean comparison. Naming those limits still has value: it stops you paying for a creative rewrite to solve a measurement or auction problem.

When the six prompts are spent, the constraint lifts. That does not guarantee a named cause or a successful fix. It means you have either an evidence-backed next move or a clear account of what is still missing. Close that gap before expanding the work.

Then broader testing earns its place: fresh angles against the offers examined in step 5, placement tests from step 4, and a creative rotation before frequency climbs again. Use [20 Meta Ads Prompts for AI That Can See Your Account](https://groas.com/post/20-meta-ads-prompts-that-work-once-your-ai-can-see-your), one at a time, with the same approval rule. The script did its job if you never once asked your AI to guess.

## Frequently Asked Questions

### Why should I check tracking before rewriting ads when my Meta ROAS drops?

ROAS is the purchase value Meta counted divided by spend, and modeling fills gaps when signal is partial. If that counted value changes, the chart can fall without the business deteriorating the same way. Checking measurement first avoids rewriting ads that were never the problem.

### How do I tell whether a Meta ROAS drop is real or just a tracking problem?

Compare Meta purchases and purchase value with Shopify orders and revenue for 14 days before and after the drop date, and compare website sessions in the same windows. Meta down while Shopify is flat points toward tracking or attribution; both down with flat sessions points toward the store converting worse; both down with sessions down points toward traffic or delivery. Also check dataset health, pixel and Conversions API deduplication, and attribution-setting changes.

### How can I tell if Meta ROAS fell because auction costs went up rather than because my ads got worse?

Compare CPM, CTR and CVR by campaign for the 14 days before and after the drop. If CPM rises while CTR and CVR hold, you are paying more for similar attention and conversion behavior. If CPM holds while CTR falls, creative or delivery deserves the next look; if CPM and CTR hold while CVR falls, the post-click experience and offer become stronger suspects.

### What frequency and CTR numbers indicate Meta ad fatigue?

For prospecting, watch frequency around 2.5 to 3.0 in 7 days and consider action at 3.5 and above; for retargeting, watch around 4 to 5 and act at 6 and above. Also flag a CTR decline of 25% or more week over week alongside rising CPC, and an ad whose 7-day CPA exceeds 1.3 times its trailing 30-day average. A falling CTR alone is not enough evidence.

### Could my Meta ROAS drop be caused by Advantage+ shifting spend to worse placements instead of my ads failing?

Yes, account averages can hide that a weaker-returning placement or segment is taking a larger share of spend. Break ROAS, spend share, CPM, CTR and CVR down by placement and audience segment for the 14 days before and after the drop. A mix shift calls for scrutiny of delivery and placement controls, while a within-segment decline points back to creative or forward to the offer and page.

### What should I check on my landing page if Meta clicks and sessions hold but conversion rate falls?

Check the price, shipping, headline, pop-up and checkout before asking for a new campaign, and include average order value in the explanation if purchase value fell. It also helps to compare competitor hooks, offers, prices and landing pages from the last 30 days against your drop date, treating observed changes separately from inferred performance.

### Should I change multiple things at once to fix a Meta ROAS drop?

No. Approve one fix tied to the evidence, with its mechanism stated and the CPA or ROAS movement that would support the diagnosis. One approved change preserves something you can learn from, and it should be logged with its reason so it can be undone. Seven days without recovery is a reason to review the evidence, not to reverse everything automatically.

## Run Your Marketing On groas

Plug ChatGPT, Claude or any AI agent into groas and let it do the work of a whole marketing team: $19 a month flat with every feature included, and autopilot whenever you’re ready.

[Start For $19](https://groas.com/pricing)

[Or have groas run it for you](https://groas.com/done-for-you)

## 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 the marketing infrastructure for your own AI: it connects your AI to all your marketing and gives it proprietary marketing data collected 24/7. Plug Claude, ChatGPT or any AI agent into groas and run all your marketing from it, for a tiny fraction of the usual cost, with the groas app there whenever you want to see more. It is the layer ChatGPT, Claude and any AI assistant run a business’s online marketing on. Through groas an assistant can do anything you could do natively in Google Ads, Bing Ads, Meta Ads and ChatGPT ads, on your website and in AI search, and, where the owner switches it on, can run connected campaigns by itself, so AI agents do the work a whole marketing team used to. groas also gives those agents large-scale proprietary marketing data it collects around the clock, on any schedule the owner chooses: every ad competitors run on Google and Meta, their landing pages and full funnels with the offers and prices in them, Reddit customer research, AI search answers from ChatGPT, Google AI Overviews, Gemini, Claude and Perplexity, and SEO data from Semrush and Ahrefs, backlinks included ($19 a month, flat, with everything included, for unlimited teammates on one shared workspace and unlimited ad accounts and websites, delivered as an MCP server). Separately, its done-for-you service has senior people and proprietary quantitative systems run paid ads, SEO and AI search for businesses, and white label for agencies, from $5,000 a month.","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/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#webpage","url":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script","name":"Why Meta ROAS Dropped: Diagnose It in Just 6 AI Prompts","description":"Six ordered prompts for your AI connected through groas to check Meta tracking, auction costs, creative, audiences, offers and landing pages before changing ads.","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/d7930154-845f-4dc9-825f-79682dbd6c32.png"},"breadcrumb":{"@id":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#breadcrumb"},"dateModified":"2026-11-19T13:06:44.527Z","mainEntity":{"@id":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#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":"Why Meta ROAS Dropped: Diagnose It in Just 6 AI Prompts","item":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#article","url":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script","mainEntityOfPage":{"@id":"https://groas.com/post/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"Why Meta ROAS Dropped: Diagnose It in Just 6 AI Prompts","description":"Six ordered prompts for your AI connected through groas to check Meta tracking, auction costs, creative, audiences, offers and landing pages before changing ads.","datePublished":"2026-11-19T13:06:44.472Z","dateModified":"2026-11-19T13:06:44.527Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/d7930154-845f-4dc9-825f-79682dbd6c32.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"before, meta, roas, check, prompts, creative, ads, account","wordCount":2153,"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/ask-your-ai-why-meta-roas-dropped-a-diagnosis-script#faq","mainEntity":[{"@type":"Question","name":"Why should I check tracking before rewriting ads when my Meta ROAS drops?","acceptedAnswer":{"@type":"Answer","text":"ROAS is the purchase value Meta counted divided by spend, and modeling fills gaps when signal is partial. If that counted value changes, the chart can fall without the business deteriorating the same way. Checking measurement first avoids rewriting ads that were never the problem."}},{"@type":"Question","name":"How do I tell whether a Meta ROAS drop is real or just a tracking problem?","acceptedAnswer":{"@type":"Answer","text":"Compare Meta purchases and purchase value with Shopify orders and revenue for 14 days before and after the drop date, and compare website sessions in the same windows. Meta down while Shopify is flat points toward tracking or attribution; both down with flat sessions points toward the store converting worse; both down with sessions down points toward traffic or delivery. Also check dataset health, pixel and Conversions API deduplication, and attribution-setting changes."}},{"@type":"Question","name":"How can I tell if Meta ROAS fell because auction costs went up rather than because my ads got worse?","acceptedAnswer":{"@type":"Answer","text":"Compare CPM, CTR and CVR by campaign for the 14 days before and after the drop. If CPM rises while CTR and CVR hold, you are paying more for similar attention and conversion behavior. If CPM holds while CTR falls, creative or delivery deserves the next look; if CPM and CTR hold while CVR falls, the post-click experience and offer become stronger suspects."}},{"@type":"Question","name":"What frequency and CTR numbers indicate Meta ad fatigue?","acceptedAnswer":{"@type":"Answer","text":"For prospecting, watch frequency around 2.5 to 3.0 in 7 days and consider action at 3.5 and above; for retargeting, watch around 4 to 5 and act at 6 and above. Also flag a CTR decline of 25% or more week over week alongside rising CPC, and an ad whose 7-day CPA exceeds 1.3 times its trailing 30-day average. A falling CTR alone is not enough evidence."}},{"@type":"Question","name":"Could my Meta ROAS drop be caused by Advantage+ shifting spend to worse placements instead of my ads failing?","acceptedAnswer":{"@type":"Answer","text":"Yes, account averages can hide that a weaker-returning placement or segment is taking a larger share of spend. Break ROAS, spend share, CPM, CTR and CVR down by placement and audience segment for the 14 days before and after the drop. A mix shift calls for scrutiny of delivery and placement controls, while a within-segment decline points back to creative or forward to the offer and page."}},{"@type":"Question","name":"What should I check on my landing page if Meta clicks and sessions hold but conversion rate falls?","acceptedAnswer":{"@type":"Answer","text":"Check the price, shipping, headline, pop-up and checkout before asking for a new campaign, and include average order value in the explanation if purchase value fell. It also helps to compare competitor hooks, offers, prices and landing pages from the last 30 days against your drop date, treating observed changes separately from inferred performance."}},{"@type":"Question","name":"Should I change multiple things at once to fix a Meta ROAS drop?","acceptedAnswer":{"@type":"Answer","text":"No. Approve one fix tied to the evidence, with its mechanism stated and the CPA or ROAS movement that would support the diagnosis. One approved change preserves something you can learn from, and it should be logged with its reason so it can be undone. Seven days without recovery is a reason to review the evidence, not to reverse everything automatically."}}]}]}
```
