Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Google Ads & PPC › Bidding & Budgets

Bidding With Few Conversions

How to bid in Google Ads when conversions are rare: consolidate campaigns, bid on a more frequent action, pool in portfolios and judge over longer windows.
Guideadvanced10 min readUpdated

Why few conversions are a bidding problem

Every Smart Bidding strategy works the same way at heart: for each auction it predicts the chance of a conversion, and its value, from the signals available, and bids accordingly. The predictions come from what the account has recorded. Google's conversion tracking documentation describes conversions as the actions a business values after someone interacts with an ad, recorded by the account's tracking, and those recorded actions are the whole of what the strategy learns from.

The Smart Bidding strategies guide states the consequence: a model with many recent conversions predicts better than one with few, and a campaign whose conversions are rare and irregular will hold a target loosely. Its status column can show a limited status for low conversion data, and the guide's table lists the responses: broaden the conversion action, pool campaigns, or run without a target longer.

Few conversions are not the same as a small budget. A kitchen installer, a commercial cleaner or a stove fitter can spend a healthy sum and still record a handful of enquiries a month, because each one is valuable and slow to arrive. The same thinness affects the advertiser as much as the model: a cost per acquisition (CPA) built on a few conversions swings widely from month to month by chance, so any judgement made on it is as shaky as the strategy's. The guide to reading small numbers explains why rates built on small counts jump around.

OptionWhat it does for the dataWhen it fitsWhat it costs
Consolidate campaignsPools conversions that were split for no control reasonSeveral campaigns share a goal, budget logic and targetSome reporting granularity
Bid on a more frequent actionGives the strategy more events to learn fromA step before the sale predicts the sale wellRisk of optimising to the wrong step
Portfolio strategyPools data across campaigns kept separate for other reasonsCampaigns share a goal and a marginOne shared target for all of them
Maximise conversions without a targetUses a thin signal without asking it to hold an averageReliable tracking, few conversionsLess control over cost per conversion
Manual or capped click biddingDoes not rely on conversion learning at allToo few or untrustworthy conversionsNo auction-time prediction

Consolidate before you automate

The first remedy costs nothing but tidiness. Many low-volume accounts are low-volume per campaign rather than in total, because the conversions are split across campaigns created by theme, by location or by habit. Each campaign's strategy then learns from its own small share.

The account structure guide sets the rule: under Smart Bidding, campaigns should be as few as the control needs allow, and what remains of granular manual-era structures is fragmentation, with ad groups and campaigns each holding a handful of conversions a month and too thin to tell whether a change helped. Its test for a campaign split is a control need: its own budget, its own target or strategy, its own location or language settings, a different conversion goal, or reporting only the campaign level can give. A split by keyword theme alone belongs at ad group level. The account audit counts campaigns with too few conversions for their bid strategy to learn, which is the list to start from.

Consolidation is not one campaign for everything. Keep separate campaigns where the economics differ, such as a high-margin service beside a low-margin one, or where a budget must be guaranteed. Merge the rest, keeping the themes as ad groups so that ads and landing pages stay relevant. The academy exercise on steering bidding through a sale and an outage for a garden building maker practises deciding which thin quotation campaigns to merge and which action the strategy should bid towards.

A more frequent action as the bidding signal

When even the consolidated campaign records too few conversions, the next lever is what counts as one. The Smart Bidding guide names the option: optimise towards a more frequent action that correlates with the real goal. For a lead business, that usually means a step earlier in the sales process than the sale itself.

The Google Ads setup for lead generation gives the practical version. Closed deals are the truest signal but often too rare to bid on, so import an earlier stage, such as a qualified or contacted lead, as the primary action, and keep closed deals as a secondary action for reporting until volume grows. The offline conversion import entry explains how outcomes recorded outside the website are uploaded against the original click, which is what lets the real outcome stay in view.

The danger is choosing a step that is frequent but uninformative. The measurement hub warns that a primary action with a high conversion rate and little value will be bought in quantity; the strategy treats whatever it is told to count as worth chasing. Three checks keep the proxy honest. Pick the step closest to the sale that still occurs often enough to help. Check, from the sales records, how often that step becomes a customer, and whether the rate is similar across campaigns and keywords; a step that converts to sales well from one campaign and badly from another will steer money the wrong way. And keep reporting the real outcome beside it, so a drift in quality shows before it costs a quarter.

Pooling campaigns in a portfolio

Some campaigns have to stay separate, for budget, location or reporting reasons, yet share a goal. A portfolio bid strategy applies one strategy and one target across them. The Smart Bidding guide describes the pooled data as the main benefit: several campaigns with modest conversion volume each learn better as one portfolio than as several separate strategies.

The price is the shared target. The portfolio allocates spend across its campaigns to hit the target overall, so one campaign can run above it while another runs below. That is right when the campaigns are economically equivalent and wrong when they are not. Group campaigns into a portfolio when they share a goal and a margin; keep them apart when they do not, and accept thinner data as the cost of a correct target. Bid limits on a portfolio are documented and generally discouraged, for the reasons the Smart Bidding guide gives.

Strategies that need less data

The choice of Bid Strategy matters more when data is thin, because strategies differ in what they ask of it.

Maximise conversions without a target asks least of a conversion-based strategy: spend the budget to get as many conversions as it can, without holding an average. The comparison of maximise conversions and target CPA notes that it works with thin history, while a target needs enough volume to hold an average reliably, and recommends maximise conversions, or pooling campaigns, before adding a target when conversions are very few.

Manual CPC, or maximise clicks with a maximum cost per click, does not rely on conversion learning at all. The comparison of manual and automated bidding lists accounts with only a few conversions in a typical month among the cases for manual bidding or an automated clicks goal with caps, and keeps manual control for conversion tracking that is broken or under review. The Smart Bidding guide describes maximise clicks with a maximum CPC as a common bridge to Smart Bidding, because it gathers data faster than manual bidding on a new campaign.

The order that usually works is: consolidate, choose the bidding signal, then pick the least demanding strategy the data supports, and move up only when the evidence does. A product on its launch day is the extreme case, with no conversions of its own at all, and the guide to paid search for a product launch applies these strategies to the launch window and says when a target can be added.

Targets on thin data

A target is a promise the strategy tries to keep on average, and on thin data the average is hard to keep and hard to read. The Smart Bidding guide's approach applies with extra force. Run without a target first, let the campaign show what it achieves at the budget, then add a target near the observed figure and tighten it in steps. Starting with an aggressive target produces very little spend, because the model finds few auctions it believes will meet it.

Thin data changes three details. Compute the observed figure over a longer window, so it rests on enough conversions to be stable. Set the first target generously, near or slightly above the observed cost, and keep it within the economic ceiling the CPA calculator gives from customer value and conversion rate. And hold it: the Smart Bidding guide lists large target changes among the triggers for a learning period, and on thin data each change takes longer to read. A target that is changed every few weeks in a low-volume campaign is never actually tested.

Judging results over longer windows

The advertiser's patience matters as much as the strategy's. Low-volume campaigns need longer windows for every judgement: whether a strategy works, whether a target is right, whether consolidation helped. Weeks are too short; in most low-volume accounts the honest unit is a month at least, and often a quarter.

Two effects make short windows worse. Conversion Lag is usually long in these accounts, because valuable purchases are considered slowly and imported outcomes arrive weeks after the click, so recent periods are always incomplete. And Sample Size is small, so a difference between two periods can be chance. Read closed periods, compare like with like, including the same season a year earlier where demand is seasonal, and judge changes at account level rather than campaign by campaign, since consolidation moves conversions between campaigns. When a strategy change does seem to have made things worse, the results dropped after a bid strategy change diagnosis lists too few conversions for the strategy to learn from among its causes and sets out how to tell it from the others. The bidding hub sets out how targets derive from margin, which is the standard every reading is judged against.

A stove installer with a few leads a month (illustrative)

Illustrative example — not a real result

A stove installer gives its bidding more to learn from

Cindermint Stoves, an invented installer of wood-burning stoves, runs three search campaigns, one each for wood burner installation, multi-fuel stoves and log burner fitting, each on target CPA with its own budget. Together they record a few quote requests a month, and all three show a limited status for low conversion data. The site also offers a free home survey, booked online; the business's records show that most installations start with a survey, and that about one survey in five becomes a sale.

The plan has three parts. The three campaigns become one, with the themes kept as ad groups. Survey bookings and quote requests both become primary conversion actions, counted once per click, and installations sold are imported from the job records as a secondary action. The merged campaign runs on maximise conversions without a target for the first quarter.

MeasureQuarter beforeQuarter after
Campaigns31
Spend£4,200£4,300
Primary conversions the strategy saw11 quote requests13 quote requests and 41 survey bookings
Installations sold, from job records47
Cost per installation sold£1,050£614

Because stove demand rises into autumn, the team also checks the same quarter a year earlier, when the old structure sold five installations. The improvement holds against both baselines, so the next quarter adds a target CPA of £85 per primary conversion, near the £80 the campaign achieved, and leaves it alone.

Common mistakes

  • Automating before consolidating. Several thin campaigns learn less than one campaign with their combined data.
  • Bidding on a frequent action that does not predict sales. The strategy buys cheap, useless conversions.
  • Dropping the real outcome from view. Import it and report it beside the proxy.
  • Pooling unlike campaigns in a portfolio. One shared target, wrong for every member.
  • An aggressive target on thin data. The campaign spends little and teaches nothing.
  • Changing the target every few weeks. On low volume, no change is ever given time to be read.
  • Judging on a week. Conversion lag and small counts make short windows noise.

Frequently asked questions

How many conversions does Smart Bidding need?

Google does not publish a minimum on the pages this guide relies on, and the site's Smart Bidding guide notes that the main strategies can be chosen with little history. What holds is that predictions improve with data: a campaign with a steady flow of recent conversions holds a target more reliably than one with a handful a month. Rather than waiting for a number, increase the data the strategy sees by consolidating campaigns, bidding on a more frequent action or pooling campaigns in a portfolio.

Should a low-volume account use manual CPC or maximise conversions?

If conversion tracking is reliable, maximise conversions without a target is usually the better start, because it needs no target and makes the most of a thin signal within the budget. Manual CPC, or maximise clicks with a maximum cost per click, still fits when conversions are too few or too untrustworthy for conversion-based learning, or when the advertiser needs a firm price ceiling. Decide campaign by campaign, after consolidating.

Is it safe to bid on a micro-conversion such as a brochure download?

Only if it reliably predicts the real outcome. A frequent action gives the strategy more to learn from, but the strategy will buy as much of it as it can, so an action that rarely leads to a customer teaches it to find cheap, useless conversions. Choose a step close to the sale, such as a booked survey or a qualified lead, check how often it becomes a customer, and keep the real outcome imported and in view as a secondary action.

Terms used on this page

Tools to try

Where this page is used

Part of these learning paths

Also related

Sources

  1. About Smart Bidding — Google Ads Help
  2. About conversion tracking — Google Ads Help

Related