Every early-stage founder eventually opens Google Ads or LinkedIn Campaign Manager, looks at the cost per click and wonders whether paid ads are a shortcut or a quick way to burn runway. Before product-market fit the stakes feel higher, because every euro spent on ads is a euro not spent on product or people.
On Hacker News, founders ask which budget, channel and strategy actually worked for an MVP, and the answers tend to be thin. Small business owners in similar threads describe spending a modest budget on Google Ads and coming away distrustful of the platform's own recommendations.
Paid ads amplify a message that already converts and rarely discover that message for you. Before product-market fit they still have a use, as a small, measured learning tool rather than a growth engine.
Know what each channel is good at
The three platforms do different jobs for a B2B company.
Google Search captures demand that already exists. Someone types a problem or a category into the search bar, and your ad appears. It is the closest thing to buying intent you can pay for, provided people actually search for what you do.
LinkedIn lets you target by job title, company and industry, which suits B2B well. In Dreamdata's 2026 LinkedIn Ads Benchmarks Report, LinkedIn showed a B2B return on ad spend of 121%, against 67% for Google Search and 51% for Meta. Those figures come from Dreamdata's own attribution of closed-won deals, so treat them as one vendor's view. The same report puts LinkedIn's average cost per click at EUR 5.98, so the platform performs but is not cheap.
Meta is built for reach. For niche B2B job roles its targeting is less precise, which fits its lower attributed return in Dreamdata's data. It can still be useful for retargeting people who have already visited your site, or for products sold to broad groups such as small business owners.
Understand the minimums before you spend
Both main B2B platforms need data to work well, and early-stage companies have very little of it.
LinkedIn's own help pages state that the minimum audience for an ad set is 300 members, but LinkedIn suggests at least 50,000 to drive results, and 300,000 for Sponsored Content and Sponsored Messaging. If your ideal customer list is a few hundred accounts, you are working at the edge of what the platform is designed for. Account-based campaigns can still work there, but expect higher costs and slower learning.
Google's automated bidding has a similar constraint. Google explains that the learning period depends mainly on the number of conversions a campaign obtains, along with the bidding strategy and the length of the conversion cycle. A B2B campaign that produces a handful of demo requests a month gives the algorithm very little to learn from. Manual CPC bidding does not depend on automated learning, so it is often the more sensible place for a small test to start.
Search costs are also moving in the wrong direction. Dreamdata reports that cost per click on non-branded Google Search terms rose by 29% while click-through rate fell by 26%.
Fix the page before you buy the traffic
Paid traffic exposes every weakness on the landing page. Unbounce's analysis of 41,000 landing pages found a median SaaS conversion rate of 3.8%, compared with 6.6% across all industries, and a median of 5.1% for SaaS traffic from Google search ads.
Your page's conversion rate has a direct effect on your cost per lead, and it is largely within your control. If your page converts poorly with the traffic you already get for free, paid traffic will not change that. It will simply make the problem more expensive.
One founder on Indie Hackers compared paid Google traffic with free community posting for a low-priced product, and the community side held up well. For a B2B product, the takeaway is that fit between message and audience matters more than the choice of channel.
Run small tests and judge them on pipeline
If you decide to test, keep the scope tight and decide in advance what you want to learn.
- Google Search: a small set of high-intent, problem-specific keywords on exact or phrase match, with manual bidding and a landing page written for those queries. You learn whether people search for your problem and which words they use.
- LinkedIn: a defined list of target accounts or a tightly filtered audience, one offer and two or three creative variations. You learn which message gets the right people to engage.
- Meta: retargeting of recent site visitors only, at least at first.
Judge every test on qualified conversations and pipeline. Clicks and cost per click say little on their own, and a cheap click from someone who will never buy is no bargain.
Plan for the longer game as well. The LinkedIn B2B Institute's 95:5 rule holds that only about 5% of B2B buyers are in market at any given time, and recommends advertising mostly to the 95% who will buy later so that your brand is remembered when they do. That is brand building, and for most early-stage companies it makes more sense once a converting message exists.
What to do this week
- List ten phrases buyers would use to describe your problem and check their search volume in Google's Keyword Planner.
- Measure your current landing page conversion rate on organic and direct traffic before spending anything.
- Count your target accounts and compare that audience size with LinkedIn's suggested minimums.
- If you run a test, fix the budget and the time period in advance, and write down the one question it should answer.
- Set up conversion tracking that records demo requests and qualified meetings, not only clicks.
Paid ads earn their place when you know who you sell to and you have a page and a message that already convert. Until then, spend small, learn quickly and put most of your effort into finding that message. Stolos Digital helps B2B teams build the ad creative, landing pages and tracking that make a paid test worth running.