Google Ads for B2B: how to build campaigns that generate qualified leads

Google Ads for B2B is rarely about winning a purchase straight from the ad – you get a lead that has to be worked before it becomes a deal. That means measurement and bidding have to be set up completely differently from B2C. Low CPC is not the goal; the right click is.

Important: the conversion does not happen on your website. A B2B deal is often decided a quarter or more after the first click, by an entire buying group – in your CRM, not in the ad. If you measure on form submissions today, you are optimising towards the wrong thing.

In this post we cover:

• How Google Ads in B2B differs from B2C
• Why you should optimise towards MQLs, SQLs and closed deals – not forms
• How to connect your CRM to Google Ads and measure all the way to revenue
• Smart Bidding, match types and negative keywords in low-volume B2B
• Consent Mode v2, GDPR and how to measure ROI when the deal takes months

We are a search marketing agency working with Google Ads and conversion optimisation every day. If you are reading this, there is a good chance you are paying for leads that never become deals – and that can be changed.

In B2C a visitor can click an ad and buy a product within minutes. B2B does not work like that. Fewer people search, clicks are more expensive and the road to a deal is long.

According to WordStream by LocaliQ’s 2026 benchmark study (13,474 US search campaigns), the average cost per click in the Business Services category is $5.87, with a conversion rate of 4.85 per cent. UK click prices cannot be read straight off US data, but the level tells you something: B2B clicks are expensive.

Here is the point, though: an expensive click can be highly profitable when a single deal is worth tens of thousands of pounds. A £6 click is nothing if it leads to a £30,000 contract.

A worked example that shows the logic: with a £2,000 monthly budget and £6 per click you get a little over 300 clicks. At a 3 per cent conversion rate that is ten leads. If one of those ten becomes a deal worth £30,000, the budget has paid for itself fifteen times over – and it was lead quality that decided it, not the cost per click.

That means low CPC can never be the goal in itself. Stop chasing cheap clicks and start chasing the right clicks – the ones from people who can become customers.

The B2B sales cycle often runs from a couple of months to more than a year – and it has got longer. In Ebsta and Pavilion’s analysis of more than three million CRM deals, sales cycles were 38 per cent longer in 2023 than in 2021.

And the decision is rarely made by one person. According to Gartner, the typical buying group in a complex B2B purchase involves 6–10 decision-makers – each doing their own research.

The consequence is simple but decisive: the conversion does not happen on your website. It happens in your CRM, several weeks later.

International benchmarks point the same way: according to Dreamdata (data from the US and Europe), it takes an average of just over six months from first touch to closed deal, and a lead will have visited you from several channels before a demo is ever booked.

If you measure on the wrong horizon, you optimise towards the wrong thing. You see a form submission today and assume it went well – but the deal is decided far later, and that is what you should be steering towards.

Hundreds of leads a month at a low cost looks fantastic – right up until you check how many of them became customers.

Take a client case from the agency NAV43: an account brought in 400 leads a month at a cost per lead of roughly $45. On paper it looks excellent. But once the advertising was connected to the CRM, it turned out that only 12 per cent reached a sales conversation and just 3 per cent became customers.

The lesson is an uncomfortable one: without signals from further down the funnel, you are training the algorithm to find more of the wrong leads, faster.

This is where the difference between micro and macro conversions starts to matter. A form submission is a micro conversion – a sign of interest, nothing more. The macro conversion is the actual deal. And in B2B you usually measure in stages: MQL (marketing qualified lead – a lead marketing has judged worth pursuing), SQL (sales qualified lead – a lead sales has taken over) and Closed Won (a won deal).

We like being concrete at Sunbird, so here is the practical advice: report on pipeline and revenue, not just on cost per lead. Lead volume is a vanity metric if it never fills your pipeline.

If you want to go deeper into getting more of your visitors to convert, conversion rate optimisation is worth a closer look – it is tightly bound up with all of this.

The single biggest lever in B2B is sending signals from further down the funnel back to Google – not form submissions, but MQLs, SQLs and closed deals.

Most advertisers stop at tracking forms. That is exactly where the problem starts.

Think of it as a feedback loop: a conversion happens, the signal reaches the algorithm, the algorithm goes off and finds more like it. The loop runs clean when the conversion is a real deal. It breaks when the conversion is a form – because a form cannot tell a qualified business buyer apart from a student doing research for an essay.

The fix is to connect your CRM (HubSpot or Salesforce, for example) to Google Ads and send the real deals back. Then the algorithm knows what a good lead looks like.

Here is how to wire the measurement up properly:

• Turn on auto-tagging in your account settings – the foundation for matching clicks to deals
• Capture the GCLID (the click ID Google attaches to every visit) in a hidden form field and store it in your CRM
• Define the funnel stages as conversions: MQL, SQL and closed deal
• Send the event back to Google Ads when a lead reaches the chosen stage – with a real or estimated value per stage: a booked demo is worth more than a downloaded guide
• Set up Enhanced Conversions for Leads (ECL) as a second matching key
• Bid towards a stage that has meaningful volume

GCLID import is now the older method – Enhanced Conversions for Leads is Google’s recommended standard. You get the best results from running both: GCLID matches where it exists, and ECL matches on hashed customer data (email, phone) when the GCLID is missing.

If you have a custom-built integration with Google Ads, it is worth double-checking. On 15 June 2026 the Google Ads API closed the old UploadClickConversions call to new integrations, and Google points to the Data Manager API as the way forward. Existing integrations keep working through a transition period, and standard CRM connectors and tag-based setups are unaffected.

Why does this matter so much? Because the failure is silent. If the upload stops working without throwing an error, Smart Bidding falls back on optimising towards whatever is still being reported – usually raw form submissions. You will not notice until lead quality slowly degrades.

Expect an improvement to show within one to two months, and a full optimisation cycle to take two to three months while closed deals accumulate.

AI bidding is only as good as the signals you feed it. Give it poor leads and it becomes an expert at finding more poor leads.

Smart Bidding lets Google set your bids automatically based on the likelihood of a conversion. The same logic applies to value-based bidding (Maximise conversion value or target ROAS), where you bid towards business value rather than just the number of conversions.

If you feed the algorithm real deals instead of forms, it filters out the rubbish itself over time. Fake form submissions and bot conversions otherwise train the system to chase exactly the wrong kind of traffic.

AI Max for Search left beta in April 2026. From September 2026, automatically created assets and campaign-level broad match are upgraded to AI Max automatically – while Google has pushed the upgrade for dynamic search ads back to February 2027. According to Google’s own data, campaigns that switch AI Max on typically see around 14 per cent more conversions or conversion value at a similar CPA or ROAS, and for campaigns that have mostly run exact and phrase match the uplift is around 27 per cent. But that assumes the measurement is in place – otherwise AI Max simply amplifies your measurement errors faster.

Rules of thumb for Smart Bidding in B2B:

• At least 15 conversions per 30 days is Google’s minimum – aim for 30 or more for stable learning
• Start with target CPA or Maximise conversions, and upload values for at least 3 weeks before you switch to value-based bidding
• Optimise towards one funnel stage at a time – several at once produces duplicated signals and weaker results
• Set target ROAS around 20 per cent below your historical average to begin with, then tighten it after the learning period

In low-volume B2B with expensive clicks, exact and phrase match protect both your budget and your lead quality. “Exact match is dead” simply does not apply here.

Search intent shifts along the buying journey. At the top, people search for problems and information (“how do we reduce churn”). Further down, searches turn supplier- and price-oriented (“b2b crm pricing”, “alternatives to X”). The money sits at the bottom.

How you should match depends on how much data you have:

• Under roughly 30 conversions a month: run exact match plus tight phrase match, with manual control and a negative list from day one.
• Over 30–50 conversions a month with strong tracking: test broad match plus Smart Bidding in a separate discovery campaign, and move the winning search terms into exact or phrase in your controlled campaigns.

One thing worth knowing: phrase match has loosened and now behaves more like old broad match used to. Word order carries less weight than it once did, so do not count on it being as precise as you are used to.

And forget SKAGs (single keyword ad groups – one ad group per keyword). With today’s matching and AI, a simpler structure performs better. Splitting the account into a hundred tiny groups is now counterproductive.

Negative keywords are the cheapest lever you have. They shut out B2C searches and irrelevant traffic that would otherwise eat budget. In WordStream’s analysis of around 15,700 accounts (2025), accounts with at least one negative keyword averaged a 13 per cent conversion rate, against 4.6 per cent for accounts with none – and a quarter of all accounts still have not added a single one. Only 12 per cent reach a Quality Score of 8 or higher.

The figure does not mean that one negative keyword triples your conversion rate – well-run accounts do many things right at once – but it does show where the well-run accounts spend their time. The most common budget thieves are job and training searches: put words like “jobs”, “vacancies”, “salary”, “training”, “course” and “free” on the negative list from day one, and write ad copy that lets the wrong audience filter itself out before the click.

Without a correct Consent Mode v2 setup, Google stops processing ad data for new visitors in the UK and the EEA – and both measurement and remarketing collapse.

Consent Mode v2 has been a requirement since March 2024, when the EU’s Digital Markets Act came into force, and Google’s consent policy covers traffic from the EEA and the UK. Without a Google-certified consent management platform (CMP) and correct consent signals, Google will not process that data for ad personalisation, remarketing and full conversion measurement. PECR and UK GDPR have required consent for non-essential cookies for far longer than that.

In a study by Advance Metrics covering 1.2 million visits to B2B sites (2023), 25.4 per cent of visitors accepted all cookies straight away, while 68.9 per cent closed or ignored the consent banner. That is a large share of your data to lose if you do not handle it properly.

Three things have become central to winning that data back while staying inside the rules:

• Conversion modelling – Google fills the gaps statistically for the visitors who did not consent
• Enhanced Conversions – hashed first-party data improves matching
• Server-side tagging – more robust data collection that is less sensitive to browser blocking

Long buying processes make remarketing a necessity. When a deal takes months, you have to be able to work the prospect across the whole cycle – in search, but also through display and YouTube, where remarketing in B2B often does more good than cold reach campaigns. Build the segments on buying signals – pricing page visits, demo pages, repeat visits – rather than on all visitors. The tools are audience segmentation: Custom Intent (audiences built on search behaviour), Customer Match (your own customer lists) and In-Market (people actively looking). But all of it assumes the consent setup works.

On 15 June 2026 Consent Mode also became the sole control for the advertising cookies the GA4 tag collects for Google Ads – the Google Signals setting no longer governs this. In practice, the ad_storage parameter decides whether the data can be used at all.

One more thing worth mentioning: AI Overviews are pushing click-through rates down on informational searches – Ahrefs measured a 34.5 per cent lower click-through rate for position 1 when an AI Overview is shown – and B2B tech is one of the sectors where they appear most often. Transactional, purchase-oriented searches are affected less. In practice that means fewer but more purchase-ready clicks at the top of the funnel – which is often an advantage rather than a loss.

The right consent setup is the difference between being able to work your prospects with remarketing and fumbling in the dark. If you need help getting measurement and consent in order, that is one of the things we work with every day – do get in touch.

Do not measure on form submissions – measure on pipeline and closed deals, and expect the numbers to mature over several months.

Cost per lead is an early indicator, not the verdict. You need it to see that the campaign is alive, but it says nothing about whether those leads become deals.

The attribution window has to match reality. If the first click often happens months before the deal closes, a 30-day window is meaningless – it cuts the measurement off in the middle of the buying process.

And do not trust GA4’s standard reports blindly. GA4 is built with ecommerce as its primary use case, where the conversion happens on site. A B2B business has to set up its own conversion events and connect the CRM feedback itself. Otherwise the default view will show you the wrong picture.

An interesting detail in the last year’s figures: according to WordStream/LocaliQ’s 2026 benchmark data (US search campaigns), cost per lead fell this year for the first time in five years, while cost per click has more than doubled over ten years. Clicks are getting more expensive, but measurement set up properly means you can still bring down the cost per genuine lead.

The two channels solve different problems. Google captures the people already searching for a solution; LinkedIn gets them to start searching.

Google Ads reaches buyers who are already in an active phase – they have a problem and they are looking for a supplier. High purchase intent, high cost per click, but lower volume. LinkedIn Ads works earlier in the process and reaches the right job roles before they have started searching, through targeting on job title, industry and company.

A concrete example: an early-stage B2B SaaS with a limited budget. Should it put everything into one channel, or split?

My answer is usually to start where the demand already exists. If people are searching for what you offer, capture them first – that is the warmest traffic you can get. LinkedIn comes in when you want to build demand and reach the people who are not searching yet.

But the strongest move is to share data between the platforms. Let Google show you which search terms convert, and use that knowledge to sharpen your LinkedIn targeting. The right channel mix in your B2B marketing means you capture both the people searching now and the people who will search in three months. And paid presence is only half of it – the organic half comes down to your social media strategy.

Here are the errors we see again and again:

• Optimising towards form submissions instead of MQLs, SQLs and closed deals
• No CRM connection, so the algorithm is trained on the wrong kind of leads
• No negative keywords, or far too few – a quarter of all accounts do not have a single one
• Broad match without enough volume or tracking
• Sending expensive clicks to the home page instead of a page built for the search
• Consent Mode v2 set up incorrectly, so measurement and remarketing are choked
• Chasing low CPC instead of profitable deals
• Trusting GA4’s standard reports for a B2B flow they were never built for
• Splitting the account into SKAGs that are now counterproductive

Plug these leaks and you get more pipeline for the same budget. You rarely need more money in the account – you need the money you already have working towards the right thing. And landing pages are the leak most often underestimated: a dedicated page that matches the search is frequently the difference between an expensive click and an enquiry.

• Optimise towards deals, not forms: send MQLs, SQLs and closed deals back to Google Ads, or the algorithm will find you the cheapest leads instead of the best ones.
• Enhanced Conversions for Leads is the standard now. GCLID import is the older method, but you get the best results from running both.
• In low-volume B2B, exact and phrase match protect both budget and lead quality. Broad match needs 30+ conversions a month, strong tracking and an aggressive negative list.
• Negative keywords are the cheapest lever there is – and a quarter of all accounts have none at all.
• Consent Mode v2 is not optional. Without it, both measurement and remarketing collapse.

It starts with the measurement. Connect your CRM to Google Ads and optimise towards MQLs, SQLs and closed deals rather than form submissions. Run exact and phrase match at low volume, build a negative list from day one, and report on pipeline rather than just the number of leads.

In B2C the goal is usually a direct purchase; in B2B it is a lead that has to be worked over time. B2B has longer sales cycles – often a quarter or more – fewer searches, higher click costs and a buying decision made by an entire buying group rather than one individual. That means measurement and bidding have to be set up completely differently.

Optimising towards form submissions with no CRM connection, having no negative keywords, running broad match without the volume or the tracking to support it, and trusting GA4’s standard reports. All of them lead to the same place: budget chasing the wrong kind of traffic.

Measure on pipeline and revenue, not on cost per lead. Match the attribution window to your real sales cycle – if a deal takes three months, a 30-day window will not cover it. Expect the real numbers to mature only after several months, once the deals have had time to close.

The purchase-ready, niche keywords at the bottom of the funnel – supplier- and price-oriented searches where the buyer is already close to a decision. Fewer but sharper keywords beat broad coverage, especially when clicks are expensive.

If you want help with your B2B Google Ads, do get in touch. We are a search specialist agency and we work with clients in 10 countries.

Get in touch and you will get a free review of your Google Ads setup.