The question every PPC manager asks at 11pm on a Tuesday: what bid should I actually be at?
The honest answer is "it depends." But it depends on so many things that the answer isn't useful until you've broken those things into formulas — five of them, in BulkSheet Pro's case. Each one answers a slightly different question. Each one is right for a slightly different moment.
This is the story of when each one fires, and why the rec engine quietly chooses one of them as the default for almost every situation.
F1: Gentle Nudge — the default
The rec engine is wired to F1. When you upload a bulk sheet and let BulkSheet Pro suggest bid changes, F1 is doing the work.
The math reads the gap between a keyword's actual ACoS and its target ACoS, expressed as a ratio of the actual ACoS itself. The ratio scales the nudge: a keyword running 2× over target gets a bigger reduction than one running 1.2× over target, and the difference is proportional. Same for under-target keywords — a keyword running at half its target gets a bigger increase than one running at three-quarters.
Then it multiplies by a sensitivity factor — 0.25 by default — which is the part the user controls.
The 0.25 default is intentional. It's a "gentle nudge" — small enough that BulkSheet Pro is sampling the auction, not retraining it. A keyword 2× over target at default sensitivity gets roughly a 25% bid reduction. Not a slash, not a full reprice. Just a meaningful, recoverable adjustment.
The PPC manager who wants bolder moves can turn sensitivity up to 0.5. The one running a fragile new account can dial it to 0.05. The math is the same; the aggressiveness is theirs to decide.
F1 has one quirk worth noting: the rec engine ignores changes smaller than 5% of the target. If a keyword is within 5% of target, the math runs but the result is discarded — there's no rec. Why: at that level, the noise in any 30-day window is bigger than the signal. Adjusting bids inside the noise floor is what makes accounts wobble.
F2: Target ACoS via RPC — the surgical reprice
F2 isn't a nudge. It's a full repricing.
It computes the actual revenue per click for a keyword and multiplies by the target ACoS to derive the bid that would exactly hit your target if revenue per click stays constant. If a keyword has $40 in sales across 20 clicks and your target ACoS is 30%, F2 says the right bid is exactly $0.60.
F2 is right when you have enough data to trust the revenue-per-click number and you want to stop hedging. It's wrong when click volume is thin or when conversion has been unstable. BulkSheet Pro will compute it — but the manager has to choose to use it. The rec engine won't apply F2 automatically.
F3: % Adjustment — the asymmetric move
F3 lets you say: "If ACoS is below this number, raise the bid by X%. If it's above, cut by Y%." X and Y can be different.
This is the formula for the manager who wants their underperformers cut hard and their performers scaled cautiously — or the other way around. The asymmetry is the value. Most automated tools want symmetric changes; reality rarely is.
F4: Max CPC — the ceiling
F4 computes the theoretical maximum bid you can pay for a keyword and still hit your target ACoS: average order value × target ACoS × conversion rate. It's not a recommendation for what to bid. It's a recommendation for what not to exceed.
F4 needs an AOV (so the keyword needs to have produced at least one order) and a real CVR. It's useful on keywords where you already know they convert but you want to defend the unit economics on the upside.
F5: Manual % Adjust — the override
F5 doesn't look at ACoS at all. You say "raise this bid by 15%" or "cut it by 10%" and that's what happens. Useful for batch operations, new-keyword seeding, manual overrides on positions you have a non-data reason to change.
Why F1 is the default
If F1 is the default, why even have the others?
Because the rec engine's job isn't to be right about every keyword. It's to be safe about every keyword.
F1 makes small, scale-aware moves that are easy to reverse if they don't work. The other four are bigger swings: surgical (F2), asymmetric (F3), defensive (F4), or arbitrary (F5). They're available when the manager has more context than the math does — but they're not safe defaults. F1 is.
Four things the rec engine knows about your bid
The PPC manager who looks at the Keywords page and sees BulkSheet Pro's bid recommendations isn't just looking at F1's output. They're looking at F1's output filtered through three other considerations.
First, negation wins. If a keyword is queued to be negated as a search term, BulkSheet Pro doesn't generate a bid recommendation for it. It would be incoherent to optimize a bid for something you're about to turn off. The rec engine sees the negation queue and suppresses bid recs for matching keywords automatically.
Second, target ACoS is hierarchical. The same rule system that powers placement recs powers bid recs. The hierarchy walks SKU rules first, then campaign rules (matched by campaign name), then keyword rules, then campaign-type overrides for Sponsored Brands or Sponsored Display, then the account default. Your branded keywords can have one target. Your new-launch campaigns can have another. Your low-margin SKUs can have a third. The math respects all of them, and most-specific wins.
Third, bid and placement are independent levers that compound. A bid of $1.00 with a Top of Search multiplier of +50% becomes a $1.50 auction price at top-of-search. If both layers move at full strength in the same direction, the cuts compound rather than add — a 20% bid cut combined with a 20% placement cut produces a 36% drop on the placement, not 40%. BulkSheet Pro handles this with a diagnostic gate that runs per campaign: by default, only one lever (bid OR placement) fires per cycle, picked based on which has the bigger gap from target. In the specific case of a multi-keyword campaign where a placement is also clearly bad, both fire — but the nudge math on each is scaled to half its normal strength so the combined movement matches what the engine intended.
Fourth, there's a CPC safeguard that prevents reckless upward moves. When the math says raise a bid (because ACoS is under target), BulkSheet Pro checks one more thing before agreeing: what's the current cost-per-click on this keyword, and what's the theoretical maximum profitable CPC at target? The maximum is computed from the keyword's conversion rate and average order value, multiplied by your target ACoS. If the current CPC already exceeds that ceiling, the keyword is overpaying on every click — and raising the bid would be the wrong move, however briefly the ACoS looks acceptable. Instead, BulkSheet Pro caps the new bid at 90% of the current one, pulling it down. The rationale: if you're already paying too much per click, an inflated ACoS won't stay inflated for long, and the next cycle's math would flip anyway. Better to walk it back now.
Confidence, again
The rec engine doesn't act on every keyword equally. Confidence scales with the same patience-budget logic as negation: account CPO × target ACoS × thresholds. A keyword's spend needs to be at least 10× its patience budget to qualify as HIGH confidence. That's the threshold for inclusion in the default Rec Actions filter.
Keywords with thin click volume — below 10 clicks total — fall to LOW confidence regardless of spend. The math is the same; what changes is whether BulkSheet Pro is willing to recommend automating the change. For low-confidence keywords, it'll show you the calculation, but it won't pre-select them.
The export
You hit Export. 184 bid changes go out. Some are small nudges on outperformers — getting pushed up so they get more impressions. Some are larger cuts on underperformers — getting walked back so they stop bleeding. None of them are reprices; F1 doesn't reprice. They're all small, recoverable, scale-aware adjustments that probe the auction without retraining it.
The month that follows shows the effect. Costs come down. Volume holds where it was earning. Volume scales where it was underleveraged. The account moves toward target — not all the way there in one cycle, but meaningfully closer. Then next month's bulk sheet runs through again, and the same engine does it again.
That's what the gentle nudge buys you: a discipline that doesn't have to be perfect on any one cycle because it gets to keep being directionally right month after month.