Paid media for B2B technology companies: the definitive guide

We have audited dozens of B2B technology paid media accounts and rebuilt many of them. The pattern is depressingly consistent. An account that looks healthy on the surface is leaking budget into auctions that will never produce pipeline. The reports look respectable because the metrics chosen to populate them were chosen to look respectable. Underneath, the cost per qualified opportunity is two to four times what it should be.

This guide is the long version of how we think about it. Most of the conclusions we hold strongly. Some of them contradict the default settings on the platforms themselves, which is precisely why so many accounts end up where they do.

Why most B2B tech paid accounts underperform

The honest answer is that platform defaults are designed to maximise platform revenue, not your pipeline. Broad match keywords, audience expansion toggles, automated bidding optimised for raw conversion volume and value rules that fire on form submissions rather than qualified opportunities all push spend into the loosest possible interpretation of intent. For a B2B technology firm with a small total addressable market that is the wrong direction.

The second issue is that most accounts have grown by accretion. Someone added a campaign to test a vertical. Someone else added a remarketing layer. A third agency added a discovery campaign that nobody watches. Nobody has gone back and asked whether the structure still represents the buyer journey. We have written about this pattern in detail in our piece on auditing a paid account that has plateaued, which is usually the symptom that brings clients to us.

The third issue is measurement. If your conversion definition is a form fill and your attribution is last click, your media plan will optimise toward whoever submits forms last. That is rarely the same group that buys. We see accounts where the most efficient channel by reported CPL is responsible for a single closed deal in eighteen months, and the worst channel by CPL has produced three quarters of pipeline. Until conversion tracking and attribution are fixed, the rest of the optimisation work is theatre.

What we do early on with any new client is freeze the account, document the current logic, then unwind the parts that do not stand up to scrutiny. The visible metrics often get worse for a quarter while the real numbers get better. Boards do not always like that, and managing the conversation is part of the job.

The channels that actually work for B2B tech

There is no universal channel mix. There are however a small number of channels that consistently produce results for B2B technology firms when run properly, and a longer list of channels that generally do not but that people keep being talked into.

The shortlist is LinkedIn Ads, Google Ads, Microsoft Ads and increasingly Reddit Ads for technical audiences. YouTube has a role in the upper funnel for product led SaaS. Meta Ads and TikTok almost never have a meaningful role for anything we work on. DV360 and The Trade Desk make sense once spend is north of fifty thousand pounds a month and you have a serious account based marketing motion.

The matrix below is the rough shape of what we see in practice across the agency book.

We have written the full breakdown by channel for B2B tech in the LinkedIn Ads playbook, our Google Ads SaaS funnel guide, the Microsoft Ads enterprise software piece and our take on Reddit Ads for B2B. Read those if you want the operational detail. The point of the matrix above is to set realistic expectations before anyone gets sold a programmatic display package they do not need.

LinkedIn Ads: the workhorse channel

For the kinds of clients we work with, MSPs, SAP partners, SaaS vendors selling to IT and finance leaders, infrastructure firms, LinkedIn is the channel that does the most heavy lifting. It is also the channel where the most money is wasted. Both things are true at once.

The reason it works is straightforward. The targeting taxonomy maps closely to how B2B firms describe their accounts. Job title, function, seniority, company size, industry and account list. There is no other major paid channel where you can run a campaign against a list of two thousand named accounts and a job title filter and expect the impressions to land where you want them to. We have written about how this varies across verticals such as MSPs, SAP, FinTech and cybersecurity because the same playbook does not apply to every category.

The reason it is wasted is that most accounts run a single audience, a single ad format and Lead Gen Forms with a generic gated asset. That gets you a CRM full of contacts who clicked because they liked the asset image, not because they are in market. We strongly prefer split testing Lead Gen Forms against Conversation Ads or website conversion campaigns for any account where lead quality matters more than lead volume, which is most of them.

The structural choices that separate good LinkedIn accounts from bad ones, in our experience, are these. First, separate ABM campaigns from broad personas, because the bidding logic is completely different. Our piece on ABM on LinkedIn goes into how we structure the account list layer. Second, treat creative as a research function, not a one off production. A handful of formats outperform the rest by an embarrassing margin and they vary by category, which is why we have written separately about Document Ads as one of the more interesting recent additions. Third, push qualified pipeline data back into LinkedIn so the optimisation has something real to chase. We cover the how in our pipeline back to LinkedIn piece.

Frequency and creative refresh

A LinkedIn campaign with a frequency of more than seven impressions per person per week starts to fatigue hard. We rotate creative every three to four weeks at the audience level and every six to eight weeks at the campaign level. The accounts that do not refresh creative end up with rising CPMs, falling CTRs and declining lead quality at the same time, which feels mysterious until you look at the rotation history.

Google Ads for SaaS, MSPs and IT services

Google Ads is where intent lives. People who type “managed service provider Manchester” or “SAP Business One implementation partner” are showing you something LinkedIn cannot. The risk is that the surrounding inventory, broad match, search partners, display network, performance max, will spend most of the budget on impressions that look nothing like that intent.

We split a typical B2B technology Google account into three structural layers. Brand defence on exact match for the company and product names, with strict negative isolation so it does not cannibalise organic. Category and competitor search on phrase match with carefully monitored search terms, the layer where most of the genuine pipeline comes from. And a small experimental layer for new themes and informational queries, ringfenced with low budget caps so it cannot bleed into the rest. Our Google Ads SaaS funnel guide has the full structure, including how we handle different funnel stages with different match types.

For MSPs and IT services firms the geography layer is more important than the match type layer, which is the opposite of what most B2B advice suggests. We have written about Google Ads for MSPs and the geo dimension because most MSP accounts are spending in metro areas they cannot service profitably and starving the boroughs they can.

The most contested decision in any Google Ads account we audit is whether to keep running paid search on branded queries when the organic listing already ranks first. Our default position is yes, but with a much smaller budget than the agency that originally set it up wanted. We argue the case in detail in branded paid when SEO ranks.

Performance Max and the question of Smart Bidding

Performance Max can work for B2B technology, but only when the conversion signal is high quality and the asset library is large enough for Google to find a useful pattern. For most of our clients, the conversion signal is too sparse and the assets are too thin. We tend to start with manual or tCPA Search campaigns and add Performance Max only after offline conversion imports are flowing reliably for several months.

Microsoft Ads: the underrated channel

Almost every account we audit is either ignoring Microsoft Ads or running it as a low priority Google Ads import. Both are mistakes. Microsoft Ads, which still includes Bing search and the Microsoft Audience Network, is disproportionately strong for enterprise software, IT decision makers, public sector buyers and anyone working in a Windows first environment. The CPCs are typically 20 to 40 per cent lower than Google for equivalent terms and the click to opportunity rate is often higher because the audience skews older, more enterprise and more likely to be the actual decision maker rather than the researcher.

We have written about this at length in Microsoft Ads for enterprise software. The short version is that you should not run it as a copy of your Google account. The match type behaviour is different, the search partner network is different and the LinkedIn profile targeting layer that Microsoft offers does not exist anywhere else. If your buyers are senior enterprise IT, ignoring this is leaving money on the table.

For one mid market enterprise software client we work with, Microsoft Ads now produces around a third of paid pipeline at roughly 60 per cent of the CPA of Google. That ratio is not unusual once you actually treat it as a first class channel.

Negative keywords and account hygiene

This is the unglamorous part of the job and the part that has the largest delta between accounts that work and accounts that do not. Negative keyword discipline is what stops a Google Ads account from quietly turning into a slush fund for “free”, “jobs”, “salary”, “course”, “tutorial” and “github” queries. We have a piece dedicated to negative keyword strategy for B2B tech because the right list is category specific and changes over time.

The rough shape of what we do every month on every account, regardless of platform:

• Pull the search terms report and segment by spend, conversions and conversion value
• Identify any term over a defined spend threshold with no conversions in the last 60 days, then categorise as add as negative, add as exact positive or watch
• Pull placement reports for display and audience network, exclude any placement above the impression threshold with no engagement
• Review demographic and device performance for any audience above the spend floor
• Check ad strength and rotation, replace anything that has been running more than eight weeks
• Reconcile reported conversions against CRM imports and flag any divergence

For accounts running Performance Max, the placement report exclusion list is now critical because the channel will otherwise spend on app inventory and YouTube channels that have no relationship to your audience.

The other half of hygiene is bidding. For low volume B2B accounts, automated bidding does not have enough data to converge. We are quite opinionated about this and have written it up in low volume B2B bidding. The short version is that under roughly thirty conversions per campaign per month we do not trust Smart Bidding and we run manual or tCPA with very narrow guardrails.

Demand generation versus lead generation: budget allocation

This is the conversation we have with every client at the start of the engagement and the one that takes the longest to resolve. The pure lead gen position says all paid spend should be measurable to a form fill or demo request. The pure demand gen position says all paid spend should be brand and category, never gated. Neither is right for any B2B technology firm we have ever worked with.

Our default split for an established B2B technology firm running a serious paid programme looks roughly like this.

These ratios shift with maturity. A firm that has only just started taking paid seriously should over index on demand capture early and earn the right to spend on demand creation later. A firm with a strong category position and an established brand can comfortably push half its budget into demand creation without seeing immediate response, because the response is showing up as elevated branded search and faster sales cycles further down.

The detailed thinking is in demand gen vs lead gen budget allocation. The reason it matters is that the wrong split for the stage you are at is the most common reason paid programmes plateau in our experience.

Retargeting without burning the brand

Retargeting is the single easiest place to do damage. The default settings on most platforms will show your ads to anyone who touched your site for ninety days at a frequency that makes them hate you. A few simple disciplines fix this.

Cap frequency at three to four impressions per week per person across the retargeting layer. Segment retargeting audiences by depth, with separate creative for high intent visits, mid intent visits and bouncers. Exclude existing customers, current opportunities and recent unqualifying disqualifications from any retargeting campaign by default. Refresh creative every three to four weeks. Put a hard time decay on the audience so anyone who has not visited in 30 days drops out, except for high intent visitors who get a longer window with different creative.

The full mechanics, including the audience structure we use across LinkedIn and Google, are in retargeting tech buyers. The reason this matters more than people think is that brand health is genuinely measurable, branded search volume, direct traffic, branded mentions, and we have seen accounts where aggressive retargeting suppressed branded search because the audience was tired of the brand by the time they were ready to buy.

Conversion tracking for long sales cycles

If your sales cycle is six to eighteen months and your conversion event is a form fill, the platform is optimising for the first thirty seconds of the journey and ignoring the next year. That is the structural reason most B2B technology paid accounts produce questionable results, and fixing it is more important than any creative or targeting work.

The chain we want is roughly this. A first party event fires when someone takes a meaningful action, ideally server side via server side tagging so it survives ad blockers and ITP. That event lands in GA4 and the CRM with a stable identifier. The CRM enriches the lead with firmographic and progression data. Stage transitions in the CRM, MQL, SQL, opportunity, closed won, are pushed back to the ad platforms via offline conversion imports as separate conversion actions. Bidding is optimised against the deepest stage that has enough volume to learn against, usually MQL or SQL rather than form fill.

We have written this up properly in conversion tracking for long sales cycles. The implementation work is unglamorous, server side GTM, Conversion Linker tags, offline conversion imports from HubSpot or Salesforce or Pardot or Marketo. None of it appears in a creative case study. All of it is the difference between an account that compounds and an account that plateaus.

Attribution models for multi touch journeys

Attribution is the area where we have the strongest opinions and the least patience for vendor pitches. Most attribution dashboards in B2B technology firms are decorative. They show last click or position based or a custom model that nobody can explain, and they get used to justify decisions that have already been made.

Our actual position is this. No single model is correct. We use a combination of platform reported conversions, GA4 data driven attribution where the data volume supports it and CRM stage data with a multi touch view that gives meaningful weight to the first touch, the opportunity creation touch and the closing touch. We treat anything in between as noise unless there is a specific touch that visibly accelerated the cycle.

The implications matter. If you measure on last click, LinkedIn looks weak. If you measure on first touch, Google looks weak. If you measure on opportunity creation touch, the picture becomes more useful but still incomplete. We work through this in attribution models for multi touch tech buying and we link it to the back end work in pipeline back to LinkedIn because attribution is only useful if the ad platforms can see the result.

The thing we never do is rely on a third party MTA tool to settle a debate. The data quality is rarely good enough and the model is rarely transparent enough.

Landing page CRO

A paid media programme is only as good as the landing experience it sends people to. We have audited accounts where the cost per qualified opportunity dropped 40 per cent in a quarter purely from rebuilding the landing pages, with no change to the media plan. That is not unusual.

The patterns we see consistently in B2B technology landing pages, and the things we change first:

• Hero clarity. The headline should describe what the product or service does in concrete terms, not what it makes the buyer feel. We have written about this in detail across our web design guide and the broader landing page CRO piece for context.
• CTA primacy. One primary action above the fold. The most common failure mode is asking for a quote when the buyer is still trying to understand what you do. We dedicated a piece to why the quote CTA is failing because it is the single most common mistake we see.
• Proof density. Logos, named case studies, technical specifics. B2B technology buyers are sceptical and the page needs to earn the form fill.
• Form length. Match the form to the offer. A demo request can be longer. A whitepaper download should not ask for budget and timeline.
• Page speed. Anything over three seconds on mobile loses a meaningful share of the click you have already paid for.

The full operating manual for paid traffic specifically is in landing page CRO for paid traffic in tech. It is the highest leverage area outside of conversion tracking. If we had to pick one thing for a client to fix first, it would usually be this rather than the media plan.

Why the quote CTA fails

A short word on this because it is the single most common pattern we see in MSP and IT services landing pages. “Get a quote” assumes the buyer has already decided. They almost never have. They want to understand fit, see comparable customers and form a view on whether you are credible before they invest in a sales conversation. A scoping call, a self serve readiness check or a content download will outperform “get a quote” by a wide margin in most categories. We unpack this further in the quote CTA piece.

How we audit a plateaued account

Most engagements with us start with an audit because the existing programme has stopped producing more pipeline as the budget rises. The audit is structured rather than impressionistic, and we have written it up at length in auditing the paid plateau. The high level shape is this.

We start with the conversion infrastructure, because there is no point auditing a media plan against a measurement system that lies. We trace the path from a click to a CRM record to a closed won deal and we look for breaks. Server side tagging, offline conversion imports, identifier consistency, deduplication. About a third of audits stop here for a quarter while we rebuild the measurement before any media decisions are made.

We then look at account structure across each platform. Are demand and capture campaigns separated. Is brand isolated. Are negative lists current. Are bidding strategies appropriate to the volume. We benchmark against what we would build if we were starting today.

We look at audience and creative health. Frequency, fatigue, rotation, format mix. We pull the search terms and placement reports for the period and quantify how much spend is going to definitively wrong matches.

We look at the landing experience and the form to opportunity conversion rate, which usually tells us more than the click to lead conversion rate.

Finally we look at the budget split across stages and channels and ask whether it matches where the firm actually is in its growth, the conversation about demand gen versus lead gen usually happens here.

The output is a prioritised list of changes with expected impact, sequenced so that measurement and structural fixes go first and the media plan adjustments follow. It is rarely glamorous and almost always works.

We have run versions of this for managed service providers, SAP partners, enterprise software vendors and IT support firms across mid market and enterprise. Companies of the type of Littlefish, Codestone, Aspire Technology Solutions and Acronyms IT Support are the kinds of clients we are designed for. Different categories, similar discipline.

The agency view, after rebuilding enough accounts to recognise the patterns, is that paid media for B2B technology is a discipline problem more than a creative one. The firms that win are the ones that get the measurement right, structure their accounts around the buyer journey rather than the platform’s defaults and refuse to optimise for vanity metrics. The ones that struggle are the ones that keep trying to find a clever creative or a new channel to fix what is fundamentally a tracking and structure problem.

If your account has stopped producing more pipeline as the budget rises, or if the reported numbers are starting to feel disconnected from what sales is closing, that is usually a sign the structural work is overdue. We are happy to take a look. Our paid media service page covers how we engage, and the contact page is the fastest way to start a conversation. The audit is paid, scoped and short, and you keep the output regardless of whether you decide to work with us afterwards. That is the only honest way to do this work.