Your ads stopped working. Here is the order to check things in

The account was fine three weeks ago. Same budget, same campaigns, same creative. Now the leads have dried up, or the sales have, and the dashboard has gone from good news to something you open with one eye shut.

The instinct is to go into the ad account and start changing things. That is usually the most expensive move available, and the account is rarely where the cause sits.

The list of suspects is well covered. Google's own troubleshooting page runs through ten reasons ads underperform, from billing and disapprovals to low bids, narrow targeting and conversion tracking (Google). WordStream offers seven fixes, Marketer.com six steps, and Jodie Minto and Anirup work through the usual Meta mistakes. Pavado and Cometly go a step further and give a diagnostic sequence. All of them are worth reading.

We agree with almost all of it. Jodie Minto puts the core point better than most: your ads probably are not the problem, the system around them is (Jodie Minto).

But nearly every sequence starts inside the ad platform and works outwards. We think it should run the other way. The list is right. The order is the problem.

Why the order matters more than the list

The cheap checks sit outside the account

Checking whether your website changed costs nothing. So does checking whether search demand dipped. Changing a bid strategy, a budget or an audience does cost something, because significant edits push a campaign back into the learning phase, where delivery is unstable while the algorithm recalibrates (Cometly). That phase typically needs 7 to 14 days and around 50 conversion events to settle (Marketer.com).

Rebuilding is worse. A fresh campaign discards its conversion history and can mean roughly two weeks of near-zero leads before it is even back where you started (Pavado).

So check in order of cost. Free and reversible first, expensive and irreversible last.

Touching the account first destroys the evidence

Every change you make inside the account is a new variable. If the real cause was a broken form on the website and you also changed the bidding, you now have two things moving and no clean way to tell which one fixed it, or which one made it worse.

The algorithm makes outside problems look like inside ones

This is the part that catches experienced people out. Smart bidding reacts to whatever signal it receives, so a problem outside the account shows up as a problem inside it.

Pavado documents a case worth knowing in full. A Google Ads account producing about 25 leads a day fell to one or two, with nothing changed in the account. What had changed was the landing page: a department manager had the call tracking number swapped for her direct line, which broke call conversion tracking (Pavado). The algorithm read the missing conversions as failure, bid lower, lost position, and every conversion type fell, not just calls. The drop arrived about four days after the change, which is why the client had already ruled it out (Pavado).

From inside the account, that was a campaign that stopped working. It was not.

And the platform will not warn you. A landing page redesign, a form change or a new offer triggers no alert in Ads Manager (Cometly).

The order we check things in

One exception first. If impressions fell to near zero overnight, go straight to the account. Something switched it off: billing, a policy disapproval, a verification check, an end date. Google lists account and billing issues first for good reason (Google), and an automated security review can limit delivery without any dashboard notification (Pavado).

For everything else, outside in.

1. Did the business number fall, or only the reported one?

Put the platform next to the business. Orders in Shopify, qualified leads in the CRM, revenue in the bank.

If Ads Manager reports 100 conversions and Shopify shows 40, that is an attribution gap, not a performance trend (Marketer.com). If the platform number fell and the business number did not, your ads may not have stopped working at all. Your measurement did. A missing Conversions API connection, events that stopped firing, or duplicated events inflating the count all distort what the platform optimises towards (Jodie Minto, Cometly).

This goes first because a broken signal corrupts every read after it. Diagnose demand or creative on bad data and you get a confident, wrong answer. It is also why we start every account we inherit here. On one German ecommerce account we rebuilt the tracking before touching the campaigns, and ROAS lifted 48% on 21% less spend (Dadek Digital).

2. Did demand change?

Then look outside the business altogether. Seasonality, a category slump, a competitor arriving at scale.

Google Trends will show whether people are simply searching less, and rising CPMs signal more competition in the auction (Marketer.com). Sometimes the market just got dearer: in US benchmark data, average search CPC rose 12% year on year to $2.96, and consumer services rose 18% (Pavado).

Pavado's test here is the cleanest we have seen. If clicks and leads fell by roughly the same percentage, the auction repriced or demand dipped. If clicks held and leads collapsed, look at tracking and the page (Pavado).

3. Did anything change after the click?

Steady click-through with falling conversions points past the ad, to the page or the offer (Anirup).

Check the website's change history, not just the account's. A redesign, a longer form, a slower mobile page, a review badge that quietly disappeared. Speed alone matters: a one-second delay can cut conversions by 7% (Marketer.com). Look back ten days, not two, because of the lag above.

For lead generation, keep going past the form. Sometimes the ads never stopped working, the follow-up did: leads still arrive, nobody answers fast enough, and the customer books the next business on the list (Pavado).

4. Did the offer go stale?

The hardest one to see, because nothing technically broke. An offer that resonated six months ago may no longer match what buyers expect or what competitors now promise, and conversion rates erode while delivery looks perfectly healthy (Cometly).

The tell is a slow slide rather than a cliff, with everything upstream checking out. The answer is rarely a discount. Compete on value and positioning, not on price (Jodie Minto).

5. Did frequency climb?

Now creative earns a look. Rising CPMs with stable click-through suggest auction competition, while falling click-through on stable CPMs suggests fatigue (Cometly). Marketer.com flags frequency above 3 to 4 over seven days for cold audiences, and 8 to 10 for retargeting (Marketer.com).

This is a common cause, and the good news is the fix is new creative, not a new campaign.

6. Only then, the campaign

If you reach this point with tracking verified, demand steady, the page unchanged, the offer competitive and frequency healthy, the cause is very likely in the account. Now the account checks earn their place.

A bid target the budget cannot support, such as a $100 cost-per-lead target on $50 a day (WordStream). Targeting that drifted, with broad match or Search Partners pulling in traffic you never chose (WordStream). Auto-applied recommendations that changed the account without anyone deciding to (WordStream). A Meta structure split so thin no ad set gets near the roughly 50 conversions a week it needs to optimise (Jodie Minto).

Make one change, write it down, and wait. Budget moves of 10 to 20% at a time keep the algorithm from resetting (Marketer.com).

What people overestimate

That creative is always the answer

Creative fatigue is real and common, which is exactly why it has become the default diagnosis. But fresh creative on top of broken tracking just teaches the algorithm the wrong lesson with new ads. A campaign with broken tracking cannot be fixed by better creative (Cometly).

That tightening the target brings costs down

It feels like discipline. Lowering a target CPA while performance is struggling tells Google to skip auctions it cannot win at that price, so delivery dries up instead of getting cheaper (Pavado). In the phone number case, the recovery involved deliberately raising the target from $50 to $100 to shorten relearning, then walking it back down (Pavado).

That there is one cause

Performance drops usually come from several factors converging (Cometly). The order still helps. Stop at the first broken thing, fix it, give it 48 hours, then carry on down the list (Pavado).

That the dashboard will tell you

The dashboard tells you something changed. It cannot tell you what, because most of the causes live in places it cannot see: your website, your CRM, your market, your phone.

Keep one change log, for the whole business

The cheapest habit that shortens every diagnosis is a single log of changes, not just inside the ad account but across the business. Logging changes is how you stop guessing at what moved performance (Marketer.com). Website edits, new phone numbers, price changes, promotions, stock-outs, a new person answering enquiries. When results drop, the cause is usually in that log, dated a few days before the drop.

Start it this week, before you need it. The next time the dashboard turns, you will be checking a list instead of changing settings.

Dadek Digital audits from the business inwards, checking tracking against real revenue before a single campaign setting is touched, so the fix lands on the actual cause. If your results have dropped off and you want to know where the problem really sits, a free audit is the fastest way to find out.