Bidding strategies for low-volume B2B keywords

The advice Google reps give on low-volume B2B accounts is almost always the same: turn on Maximise Conversions, give it more data, give it more time. The reality, in our experience, is that smart bidding will quietly strangle accounts running on twenty conversions a month. The algorithm needs more signal than these accounts produce, and starving for data, it makes worse and worse decisions until the cost-per-lead doubles and someone calls us in.

There is a real argument for smart bidding in B2B, but it depends entirely on signal volume and signal quality. Below is how we actually approach bidding on the low-volume keyword sets that most B2B tech firms run, where a “good month” might be 30 leads across the whole account.

What counts as low volume

Smart bidding strategies in Google Ads have published thresholds that the platform will quote at you (50 conversions in 30 days for Target CPA, for example). Those numbers are not really the problem. The problem is variance. A B2B account with 50 conversions a month, where 30 are from the brand campaign and 20 are scattered across a dozen non-brand campaigns, is functionally low volume at the campaign level even if it crosses the threshold at the account level.

We treat any campaign producing fewer than 15 to 20 conversions a month as low volume for bidding purposes, regardless of what the platform recommends. Below that, smart bidding makes decisions on noise.

Start with manual CPC, deliberately

For genuinely low-volume campaigns, we usually start with manual CPC. It sounds dated. It works. The reason is straightforward: when conversion data is thin, you do not want the algorithm allocating budget on the basis of three conversions, two of which were junk. You want a human reading the search terms, the device split, the day-parting and adjusting bids on logic rather than statistical confidence the data does not have.

Manual CPC also forces you to take negative keywords seriously. A campaign on smart bidding rewards you for ignoring search terms because the algorithm is “learning”. A campaign on manual CPC punishes you immediately when the wrong searches eat the budget. We work through that discipline in our negative keyword strategy for tech piece, which is the companion to this one for most low-volume accounts.

Move to enhanced CPC, not Maximise Conversions

When you have enough signal that pure manual feels coarse but not enough for full smart bidding, enhanced CPC is the bridge. It lets you set the bid floor manually and gives Google room to adjust within a defined range based on conversion likelihood. We have found this works well for non-brand B2B campaigns producing 8 to 20 conversions a month per campaign.

The mistake we see is jumping straight from manual CPC to Maximise Conversions because someone in a Google call recommended it. With low volume, that is the moment the cost-per-lead spikes, because the algorithm is now confidently chasing patterns that are not really there.

Target CPA: only with good signal

Target CPA can work in B2B at lower volumes than the platform’s stated thresholds, but only if the conversion signal is clean. That means three things in our experience. First, the conversion has to be a real conversion. Counting newsletter signups, demo views and contact-page visits as conversions makes the data noisier and the bidding worse, not better. Second, conversion lag has to be short enough that the algorithm sees the result inside its learning window. Third, the conversion value should be reasonably consistent.

For SaaS firms with cleaner signal (demo requests with reasonable lead quality, plus offline conversion imports for SQLs), Target CPA at the campaign level can outperform manual once volume is at 12 to 15 conversions a month and trending up. For MSP, ERP and consultancy accounts where every “lead” is a different shape, we are often more cautious.

Portfolio bid strategies for thin campaigns

When individual campaigns are too thin for their own bid strategy but the account in aggregate has reasonable volume, portfolio bid strategies are useful. You group several non-brand campaigns under one shared Target CPA, and Google bids across the pool rather than per campaign.

This works particularly well for low-volume B2B accounts running multiple service lines. Three campaigns each doing 5 conversions a month will not learn individually, but a portfolio doing 15 in aggregate will. The cost is loss of granular control, so we usually only group campaigns that share creative direction and audience intent.

Offline conversion imports change the maths

The single biggest improvement we make on low-volume B2B accounts is feeding offline conversion data back into Google Ads. The default tracking on most accounts treats a form fill as the end of the journey. The reality is that the form fill is the start, and the meaningful event is the SQL or opportunity created in HubSpot or Salesforce two to six weeks later.

When you import that downstream signal, two things happen. First, the bidding algorithm starts optimising against quality, not quantity. We have seen the cost-per-lead actually rise on accounts after this change while the cost-per-opportunity fell by 40 per cent. Second, you can finally use Target ROAS or value-based bidding, because each conversion can carry a value that reflects deal stage or weighted opportunity size.

Our conversion tracking guide for long B2B sales cycles walks through how we set this up, and the broader attribution models for multi-touch B2B piece explains how we use the resulting data to defend budget. For low-volume accounts, this is the change that unlocks smart bidding eventually being worth running.

Branded campaigns are a separate animal

Brand search behaves differently. Volume is usually higher, intent is high and the cost-per-click should be very low. We almost always run brand on its own bidding strategy, separate from non-brand, and we usually run it on Target Impression Share rather than Target CPA. The objective is to defend the SERP, not to optimise cost per conversion against an audience that was already going to convert.

Whether to run branded paid at all when SEO already ranks is a separate question. We worked through that in detail in our branded paid search when SEO ranks piece, which has more on the audit logic.

A practical sequence

For most low-volume B2B accounts we take over, the bidding migration looks like this:

1. Audit conversion actions, remove junk, set up offline conversion imports (week 1 to 3)
2. Manual CPC across non-brand for the first 30 days, intensive negatives and bid management
3. Enhanced CPC on the campaigns that develop volume, manual CPC on the rest
4. Portfolio Target CPA across non-brand once aggregate volume is reliable
5. Per-campaign Target CPA, then Target ROAS once offline values are flowing cleanly

That is usually a three to six month migration, not a two-week project. The mistake is doing it all in week one because the platform suggests it.

When smart bidding genuinely is wrong

Some B2B accounts will never produce enough signal for smart bidding to outperform a thoughtful manual setup. A specialist consultancy bidding on a dozen niche terms with three conversions a month is one of those. In that case, the right answer is to keep manual CPC, get very disciplined on match types and negatives and stop trying to force the algorithm. There is no prize for being on the latest bid strategy.

If you are managing a low-volume B2B account and the cost-per-lead has been creeping up since you switched to smart bidding, the answer is rarely “give it more time”. If you’d like a second opinion on attribution or budget split, drop us a line. We also cover the broader account structure on our paid media service page, and the keyword side in long-tail keywords for MSPs.