Is the Right Bidding Strategy Running? A Bidding Audit

A lot of accounts are held back not by bad keywords or weak ads, but by the bidding. Either the wrong strategy is running, or the right one is set up wrong, and in both cases the account quietly underperforms while everything else looks fine. Bidding is one of the highest-leverage things an audit checks, because a single wrong setting can cap an entire account. Here is how to audit it and what to look for.

Is the strategy even the right one for the goal?

Start with the basics: does the bidding strategy match what the account is trying to do? An ecommerce account chasing profit should usually be on a value-based strategy like target ROAS, not maximise clicks, which just buys traffic. A lead gen account should be driving target cost per action, not impression share, which buys visibility whether or not it converts.

A surprising number of accounts are on a strategy that does not fit the goal at all, often a leftover from an earlier setup nobody revisited. Maximise clicks running on an account with plenty of conversion data is leaving the machine's best tools switched off. The first audit question is simple: is this strategy even trying to optimise for the thing that matters?

Is the target strangling the account?

The most common bidding fault is a target set too aggressively. A target ROAS set too high, or a target cost per action set too low, tells Google to only bid when it is very confident, so the account starves itself of volume. Impression share collapses, spend drops, and the account looks efficient while leaving a pile of profitable volume on the table.

The tell is high efficiency with falling or capped volume. If the ROAS looks great but impression share is low and spend will not grow, the target is probably too tight. The opposite fault, a target too loose, spends freely at a poor return. In an audit, the target is the single number most worth pressure-testing, because getting it wrong in either direction costs a lot.

Is it being left alone long enough to learn?

Smart Bidding needs a stable target and time to learn. A common fault an audit finds is a strategy that is changed too often, the target nudged every few days, the strategy switched every few weeks, so the algorithm never settles and never performs.

Every significant change restarts the learning, and an account in permanent learning mode never reaches its potential. If the change history shows constant tinkering, that is itself the problem. Part of a bidding audit is checking not just the current settings, but how often they have been disturbed. Sometimes the fix is to pick a sensible target and leave it alone.

Is it being fed a good signal?

A bidding strategy is only as good as the data it optimises towards. Target ROAS chasing raw revenue, or target cost per action chasing conversions that are not all equal, will optimise hard towards the wrong thing. The strategy can be perfect and still produce poor results if the signal underneath is weak.

So a bidding audit is partly a tracking audit. Is the conversion data clean? Is value being passed, not just counts? Is that value close to profit rather than raw revenue? The best bidding setup in the world cannot save an account feeding it a bad number, which is why the signal is as much a part of the bidding audit as the settings themselves.

What a bidding audit comes down to

Four questions do most of the work. Is the strategy right for the goal? Is the target set at a level that drives volume as well as efficiency? Has it been left alone long enough to learn? And is it being fed a signal worth optimising towards? Run those checks and you catch the faults that quietly cap an account, the ones no amount of keyword or ad work can fix, because they sit a level above all of it.

FAQs

What does a bidding strategy audit check? Whether the strategy fits the account's goal, whether the target is set at a sensible level, whether it has been left stable long enough to learn, and whether the conversion signal it optimises towards is clean and value-based.

What's the most common bidding mistake? A target set too aggressively. A target ROAS too high or a target cost per action too low makes the account starve itself of volume, so it looks efficient while leaving profitable volume unbought.

How do I know if my target is too high? Look for high efficiency with low or capped impression share and spend that will not grow. Strong ROAS but stalled volume usually means the target is too tight.

Why does changing the bidding too often hurt? Every significant change restarts Smart Bidding's learning. An account whose target or strategy is constantly adjusted never settles, so it never reaches its potential. Stability matters.

Can the bidding strategy be fine but still underperform? Yes. If the conversion data is wrong or the signal is raw revenue rather than value or profit, even a well-chosen strategy optimises towards the wrong thing. The signal is part of the bidding audit.