LinkedIn Ads for B2B Startups: When to Use Them and How to Start

  1. Blog
  2. Demand & Channels

By Saroj Jha · October 4, 2026 · 10 min read

LinkedIn ads fit when your buyers are defined by their job and their company, and one customer is worth enough to pay auction prices. Start with a small, capped test aimed at a named list of accounts, and judge it by qualified pipeline, not clicks.

In three numbers

  • $10 a day: LinkedIn’s minimum budget for any ad format, and $100 for a new campaign’s lifetime[1]
  • 300 matched members: the smallest audience a LinkedIn campaign can run with[2]
  • 5% of B2B buyers are in the market in a given quarter, by the Ehrenberg-Bass rule of thumb[3]

LinkedIn is the main ad platform where you can target by job title, function and employer at scale, which is why B2B founders reach for it first. It is also easy to spend a lot on it and learn little.

Below are the five steps I would work through with a founder before and during a first LinkedIn campaign. One example company runs through all of them.

Running case: Example Co. Fictional

Example Co. is a seed-stage B2B SaaS startup whose software automates the month-end close for finance teams at companies with 50 to 500 employees. The controller is the champion and daily user, the CFO or VP Finance signs, and IT or security reviews the purchase; sometimes an outside accountant has a say. It charges about $1,500 a month (about $18,000 a year), its sales cycle is about 45 days, and its team of eight wants 30 new customers in the next 12 months. The founders already post on LinkedIn and want to know whether to pay for ads. Example Co. and all its numbers are illustrative.

Step 1: Check that LinkedIn fits your buyer

LinkedIn ads make sense when three things are true:

  1. Your buyer is defined by role and company. “Heads of finance at software companies with 50 to 500 staff” is a LinkedIn audience; “people who want to save money” is not.
  2. A customer is worth a lot. LinkedIn sells ads “through the ad auction”, charges only for the event your campaign objective pays for (a click to your site, for a website-visits campaign) and publishes no typical costs[4]. I know of no independent cost benchmark with a published method, so plan as if clicks will be expensive until your own test shows otherwise. That is why the channel suits higher deal values.
  3. You have something worth their attention. A useful guide, data or a point of view. An ad that only says “book a demo” asks a lot of buyers who aren’t ready.
Figure 1 Example Co. answers yes to all three checks, so it runs a small test
  1. 1 Is your buyer defined by role and company? Example Co.: yes. Controllers and CFOs at companies with 50 to 500 employees Yes
  2. 2 Is one customer worth paying auction prices to win? Example Co.: yes. About $18,000 a year per customer Yes
  3. 3 Do you have something worth their attention? Example Co.: yes. A month-end close checklist for controllers Yes

Three yes answers: run a small, capped test (Step 2)

Any answer no: fix that first, or put the money into another channel

My three checks, from Step 1. Example Co. and its answers are fictional.

Example Co., step 1 Fictional

All three answers are yes. Its buyers are controllers and CFOs at companies of a known size. One customer is worth about $18,000 a year. And the team has a month-end close checklist for controllers that it can offer as a genuinely useful guide (a fictional asset).

If you are still testing which channels work, my post on which marketing channels to start with helps you compare LinkedIn with the alternatives.

Do this

Write down your three answers. If any is “no”, fix that first or put the money into another channel.

Step 2: Set a small test budget and a stop rule

LinkedIn’s minimum daily budget is $10 for any ad format, and a new campaign’s lifetime budget must be at least $100; LinkedIn suggests new advertisers start at $25 a day[1].

Plan a first test that runs long enough, usually several weeks, to see how many qualified conversations it produces, and decide in advance what result would justify more spend. The Paid Media Budget Allocator helps you set the test budget against other channels, and the Unit Economics Calculator tells you what a customer is worth paying for.

Example Co., step 2 Fictional

The team assumes it can spend up to $9,000 to win one customer, half of first-year revenue. It sets $60 a day for eight weeks, about $3,360 in total. Its stop rule: fewer than three qualified opportunities by the end means stop or change the offer. All of these numbers are fictional.

Figure 2 Example Co.’s whole test costs less than it would pay to win one customer

Example Co.’s numbers Fictional

$18,000 A year from one customer Half $9,000 Most it will pay to win one customer Test $3,360 $60 a day for 56 days: 37% of $9,000 Judge 3+ Qualified opportunities by week 8, or it stops

A day’s budget, drawn to scale: LinkedIn’s floor against Example Co.’s plan

  • LinkedIn’s minimum daily budgetAny ad format $10 a day
  • LinkedIn’s suggested startFor a new advertiser $25 a day
  • Example Co., each campaignTwo campaigns, split roughly evenly About $30 a day
  • Example Co., in totalFictional $60 a day

A new campaign’s lifetime budget must be at least $100. Example Co.’s $3,360 clears it, and each of its two campaigns stays above the $10 daily minimum.

Show the numbers as a table
Number Value Source
Value of one customer, a year $18,000 Fictional
Most Example Co. will pay to win one customer $9,000 Fictional
Test budget: $60 a day for 56 days $3,360 Fictional
Stop rule: qualified opportunities needed by week 8 Fictional
LinkedIn’s minimum daily budget, any format $10 LinkedIn
LinkedIn’s minimum lifetime budget, new campaign $100 LinkedIn
LinkedIn’s suggested start for a new advertiser $25 LinkedIn

LinkedIn’s budget minimums and suggested start from its budget guidance, checked October 14, 2026. Example Co.’s customer value, budget and stop rule are fictional, from Step 2 of the running case. Bars drawn to scale.

Do this

Write the budget, the length and the stop rule on one line before the campaign starts, so the result cannot move the goalposts.

Step 3: Target a named account list and the buying group

Start narrow and specific:

  • A named account list. Upload the companies you want to reach, for example from your ICP Fit Scorer output. LinkedIn needs at least 300 matched members to run a campaign, and a contact list upload can hold up to 300,000 email addresses[2]. For a company list, LinkedIn suggests at least 1,000 companies, and turning a list into an audience takes 24 to 48 hours[5].
  • Job titles or functions and seniority, combined with company size and industry[6].
  • Retargeting people who visited your site or engaged with your ads. The retargeting post covers when this helps.

Don’t stack so many filters that the audience becomes tiny. LinkedIn’s rule of thumb is two or three targeting facets at most and, for Sponsored Content, an audience above 50,000[7]. A named account list will be far smaller. I accept that trade for a first test, as long as the list clears the 300-member minimum. The ABM Playbook covers building and prioritizing an account list.

Figure 3 A named list needs only 300 matched members to run, far below LinkedIn’s 50,000 rule of thumb
  1. Upload 1,000+ companies: LinkedIn’s suggested size for a company list A contact list file holds up to 300,000 email addresses
  2. Match 24–48 h for LinkedIn to turn an uploaded list into an audience A replacement list pauses the campaign until it is matched
  3. Run 300 matched members: the smallest audience a campaign can use Members who opt out of third-party data drop out of the count
  4. Scale 50,000+ people: LinkedIn’s rule of thumb for a Sponsored Content audience Built from two or three targeting facets at most

Example Co. Fictional

Uploads about 1,500 companies: above the 1,000 LinkedIn suggests. Its two finance audiences inside those companies will be far smaller than 50,000 people, a trade it accepts for a test aimed at named accounts.

Show the numbers as a table
Number What it is Source
1,000 companies Suggested minimum size for a company list Account Based Targeting
300,000 Most email addresses in one contact list file Matched Audiences FAQ
24 to 48 hours Time to turn an uploaded list into an audience Account Based Targeting
300 matched members Smallest audience a campaign can use Matched Audiences FAQ
50,000 Rule of thumb for a Sponsored Content audience when starting out Ad targeting best practices
2 to 3 Most targeting facets LinkedIn recommends adding Ad targeting best practices

LinkedIn Marketing API documentation: Matched Audiences FAQ (updated August 17, 2026) and Account Based Targeting (updated July 29, 2026); LinkedIn’s ad targeting best practices, checked October 14, 2026. Example Co.’s list is fictional.

Then think in people, not only companies. In my view, each person in the buying group needs a different message, and some are better reached by your sales team than by an ad.

Example Co., step 3 Fictional

It uploads a list of about 1,500 companies with 50 to 500 employees from its ICP Fit Scorer output (a fictional list). Inside that list it builds two audiences: controllers, the champions, and finance leaders, the people who sign. It runs no ads to IT and security; it answers their questions during the sales process instead.

Figure 4 Example Co. advertises to two of the three people who decide, and answers the third through sales

Named account list About 1,500 companies with 50 to 500 employees

Person What reaches them
  • Controller Champion and daily user — Audience A: controllers Document ad with a lead form: the month-end close checklist
  • CFO or VP Finance Signs the purchase — Audience B: finance leaders Thought Leader Ads from the founder’s best posts about slow closes
  • IT and security Reviews before purchase — No ads Security answers shared during the sales process

Example Co.’s plan from Step 3 of the running case; fictional.

Do this

List each person in your buying group and decide, for each one, whether an ad should reach them and with what.

Free template

Turn each buyer into a LinkedIn audience

The free Paid Audience Spec Builder turns a persona into LinkedIn targeting for each person in the buying group, and checks that each audience clears LinkedIn’s minimum before you spend.

Get the template

Step 4: Split spend between brand and lead capture

LinkedIn offers several formats. These four cover most startup needs:

Table 1. Four LinkedIn ad formats that cover most startup needs
Format What it is Best for
Document ads A PDF, Word or PowerPoint file people read in the feed (up to 300 pages; LinkedIn suggests under 10)[8]; can be gated with a lead form[9] Sharing a guide or research; capturing leads from it
Thought Leader Ads A person’s post, such as the founder’s, an employee’s or a customer’s, promoted by your company with the author’s approval[10] Building trust through a person, not a logo
Lead Gen Forms A form pre-filled from the member’s profile, with up to 12 fields (LinkedIn suggests three or four)[11] Capturing leads without a landing page
Single image and video ads Standard sponsored posts[12] Simple messages and awareness

For many startups, promoting the founder’s best posts as Thought Leader Ads is the most natural start. The Founder-Led LinkedIn Playbook covers building that organic base first. Research on one growing B2B service firm’s LinkedIn posts over 106 weeks found that growth in followers and website visits was linked to higher sales revenue[13]. It is one firm and organic posts, not ads, but it supports building an audience rather than only buying clicks.

Should you run brand ads or lead generation ads? Both, and more brand than most startups expect. Three pieces of research, all produced with LinkedIn’s B2B Institute[14], point the same way:

  • Analysis of the IPA effectiveness databank found B2B efficiency peaked when about 46% of budget went to brand and 54% to activation[15]. The authors say the ratio “should not be followed too precisely”: the B2B sample was small, fewer than 50 cases, and LinkedIn’s summary page rounds the split to 50/50[16].
  • The Ehrenberg-Bass Institute’s “95-5 rule” estimates that only about 5% of B2B buyers are in the market in a given quarter[3]. The author calls it a heuristic, not a precise rule.
  • A survey of business insurance buyers found that for each additional buying situation a customer linked to a brand, their stated chance of switching away fell by about 5%, in 16 of 17 insurance products studied[17].

The practical point: lead forms catch the few buyers in market now; brand formats make sure the rest remember you when they are ready. Judge the two with different measures.

Figure 5 About 95 in 100 B2B buyers are not in the market this quarter, so a brand lane runs beside the lead lane
  • In the market now: about 5 in 100 Not in the market: about 95 in 100
  • For the 95 — Brand lane — For the 5 — Lead lane
  • Share of budget, brand and activation — Binet and Field B2B efficiency peaked here; fewer than 50 cases — Example Co.’s test Split roughly evenly; fictional
Show the numbers as a table
Number Share Source
B2B buyers in the market in a given quarter About 5% Dawes (2021), heuristic
B2B buyers not in the market About 95% Dawes (2021), heuristic
Budget to brand where B2B efficiency peaked 46% Binet and Field (2019)
Budget to activation where B2B efficiency peaked 54% Binet and Field (2019)
Example Co.’s test, brand lane and lead lane Roughly 50% each Fictional

Dawes (2021), Ehrenberg-Bass Institute: a heuristic, not a precise rule. Binet and Field (2019), IPA Databank B2B cases, 1998 to 2018; the authors say the ratio “should not be followed too precisely”. The two lanes are my framing; Example Co.’s split is fictional. Dots and bars drawn to scale.

Example Co., step 4 Fictional

For controllers, it runs a document ad with the month-end close checklist, gated with a lead form. For finance leaders, it promotes the founder’s three best posts about slow closes as Thought Leader Ads. It splits the test budget roughly evenly between the two lanes, a fictional choice loosely following the research above.

What LinkedIn’s owner tells investors

Microsoft owns LinkedIn, and its latest annual report, for the year to 30 June 2026, describes the business in plain terms. This is what the filing says, not an account of how LinkedIn’s ad auction works or how Microsoft makes its choices internally.

The filing says: “LinkedIn connects the world’s professionals to make them more productive and successful and transforms the way companies hire, market, sell, and learn.” It names four business lines: “Talent Solutions, Marketing Solutions, Premium Subscriptions, and Sales Solutions.” Marketing Solutions is the line tied to sponsored content. On what drives revenue, it says: “LinkedIn revenue is mainly affected by demand from enterprises and professionals for subscriptions to Talent Solutions, Sales Solutions, and Premium Subscriptions offerings, as well as member engagement and the quality of the sponsored content delivered to those members to drive Marketing Solutions”[18].

Figure 6 LinkedIn’s owner ties its ad business to member engagement and the quality of sponsored content

Microsoft Corporation, Form 10-K Fiscal year ended 30 June 2026, LinkedIn

“LinkedIn connects the world’s professionals to make them more productive and successful and transforms the way companies hire, market, sell, and learn.”

The four paid lines the filing names

  • Talent Solutions
  • Marketing SolutionsSponsored content
  • Premium Subscriptions
  • Sales Solutions

What the filing says drives Marketing Solutions

Member engagement “increase LinkedIn member engagement on the platform”

+

Content quality “the quality of the sponsored content delivered to those members”

Marketing Solutions LinkedIn’s ad business

My reading. Make each ad something a finance leader would choose to read.

What the filing does not say. How LinkedIn’s auction ranks ads, so it is no promise of cheaper clicks.

Microsoft Corporation, annual report on Form 10-K for the fiscal year ended 30 June 2026, Item 1 (LinkedIn), quoted exactly. What the filing says, not how LinkedIn’s auction works. The two reading notes are mine.

In my view, the lesson for a startup is simple. By the owner’s own account, LinkedIn’s advertising business depends on members engaging with sponsored content, so an ad worth reading fits the way the platform describes itself. The filing does not say how the auction ranks ads, so do not read it as a promise of cheaper clicks.

Do this

Start with one brand format and one lead-capture format, and give each its own success measure.

Step 5: Measure by pipeline and a control group

Don’t rely on platform-reported results alone. A study of 663 advertising experiments on Facebook found that methods without a control group overstated the true effect several times over: for lower-funnel outcomes, the median lift in experiments was 5%, while the two non-experimental methods tested put it at 24% and 64%[19]. LinkedIn’s Revenue Attribution Report uses your CRM data to “measure the impact your marketing efforts have on sales metrics, such as revenue and pipeline”[20]; without a control group, that shows association, not cause.

Instead:

  • Track pipeline, not clicks. Record the source of each opportunity in your CRM and review qualified pipeline and revenue from LinkedIn every month. LinkedIn’s Insight Tag tracks conversions on your website[21], and its Conversions API can send offline events, such as a lead your sales team qualifies, so campaigns optimize for qualified leads rather than form fills[22].
  • Use tests with a control group. For pipeline, run your own holdout test. LinkedIn also offers A/B tests, which compare two campaigns, and Brand Lift studies, which survey a test group and a control group that is not shown your brand ads, to measure brand perception[23].
  • Ask new customers how they heard of you. The Self-Reported Attribution Coder helps you analyze the answers.

Example Co., step 5 Fictional

Before launch, it holds out a random 300 of its 1,500 accounts (20%) from all LinkedIn ads. After eight weeks the lead form has produced 38 leads and 5 qualified opportunities, enough to pass its stop rule, but no closed deals yet, which is expected with a 45-day sales cycle. It keeps the holdout for a full quarter and compares qualified opportunities per 100 accounts in each group before adding budget. All numbers are fictional.

Figure 7 For outcomes closest to a sale, experiments found a 5% lift; two methods without a control group found 24% and 64%

Median lift across 663 Facebook ad experiments, by method

  • Randomized experiment
  • Double machine learning
  • Propensity score matching
  • Upper funnel
  • Middle funnel
  • Lower funnel: purchases, registrations The outcome closest to pipeline

Example Co.’s holdout test Fictional

  • Ads on: about 1,200 accounts (80%)
  • Held out, no ads: about 300 (20%)
  • Each square is 100 of its 1,500 accounts
  1. Week 8 38 leads and 5 qualified opportunities: it passes the stop rule of 3. No closed deals yet, as expected with a 45-day sales cycle.
  2. After a quarter Compare qualified opportunities per 100 accounts in each group; the gap is the lift. Only then add budget.
Show the numbers as a table
Outcome Randomized experiment Double machine learning Propensity score matching
Upper funnel 29% 83% 173%
Middle funnel 18% 58% 176%
Lower funnel: purchases, registrations 5% 24% 64%
Example Co.’s holdout (fictional) Value
Accounts with ads on About 1,200 (80%)
Accounts held out About 300 (20%)
Leads after 8 weeks
Qualified opportunities after 8 weeks

Gordon, Moakler and Zettelmeyer (2023), Marketing Science: median lifts as reported in the published abstract; the two methods without a control group are double/debiased machine learning and stratified propensity score matching. Example Co.’s holdout and results are fictional. Bars and squares drawn to scale.

My post on incrementality testing for startups covers how to run a holdout test.

Do this

Tag every opportunity’s source in your CRM from day one, and set up the holdout before the first ad runs.

What changes by stage?

  • Pre-seed — Mostly organic: the founder’s posts. Promote the best ones as Thought Leader Ads on a small budget to a named account list.
  • Seed — Add document ads with a genuinely useful asset and lead forms. Measure qualified pipeline per dollar against your other channels.
  • Series A–B — Run brand and lead campaigns separately, add a control-group test each quarter, and tie spend to pipeline targets.

What mistakes should I avoid?

  • “Book a demo” as the only ad. Most of your audience is not ready.
  • Judging by cost per lead. Cheap leads that never become pipeline cost the most.
  • Audiences that are too small or too broad. Aim for a defined list of the right companies and roles.
  • Trusting platform attribution. Use your CRM and control-group tests.
  • Stopping brand ads after a month. Brand effects take time to show.

Related on AAJ

Articles

Playbooks

Templates and tools

References

The superscript numbers in the article match this list.

LinkedIn’s product pages and documentation change often; every LinkedIn and Microsoft page above was checked on October 14, 2026.

  1. LinkedIn Marketing Solutions. “Campaign budgets for your LinkedIn ads.” Minimum daily and lifetime budgets and the suggested start for new advertisers. business.linkedin.com
  2. LinkedIn Marketing API documentation. “Matched Audiences FAQ”, updated August 17, 2026. Microsoft Learn. learn.microsoft.com
  3. Dawes, J. (2021). “Advertising effectiveness and the 95-5 rule: most B2B buyers are not in the market right now.” Ehrenberg-Bass Institute, B2B report, May 2021. marketingscience.info
  4. LinkedIn Marketing Solutions. “LinkedIn advertising costs and pricing.” Ads are sold through the ad auction and billed on the event the campaign objective sets. business.linkedin.com
  5. LinkedIn Marketing API documentation. “Account Based Targeting (ABM)”, updated July 29, 2026. Microsoft Learn. learn.microsoft.com
  6. LinkedIn Marketing API documentation. “Targeting Criteria Facet URNs”, updated February 27, 2026. Job function, title, seniority, company size and industry facets. Microsoft Learn. learn.microsoft.com
  7. LinkedIn Marketing Solutions. “Best practices for ad targeting.” Audience size and the number of targeting facets when starting out. business.linkedin.com
  8. LinkedIn Marketing Solutions. “Document Ads specs.” File types, file size and page limits. business.linkedin.com
  9. LinkedIn Marketing Solutions. “Document Ads.” Gating a document with a Lead Gen Form. business.linkedin.com
  10. LinkedIn Marketing Solutions. “Thought Leader Ads.” Promoting posts by employees, industry experts, customers or creators with the author’s approval. business.linkedin.com
  11. LinkedIn Marketing Solutions. “Lead Gen Forms specs.” Up to 12 fields; three or four recommended. business.linkedin.com
  12. LinkedIn Marketing Solutions. “Objective-based advertising.” Campaign objectives and the ad formats available for each, including single image and video ads. business.linkedin.com
  13. Mora Cortez, R., Johnston, W. J. and Ghosh Dastidar, A. (2023). “Managing the content of LinkedIn posts: Influence on B2B customer engagement and sales?” Journal of Business Research, 155, 113388. doi:10.1016/j.jbusres.2022.113388
  14. Weinberg, P. and Lombardo, J. (2021). “The 95:5 rule is the new 60:40 rule.” Marketing Week, 2 September 2021. The heads of The B2B Institute describe the 95:5 rule as “our new research with Professor John Dawes”. marketingweek.com
  15. Binet, L. and Field, P. (2019). The 5 Principles of Growth in B2B Marketing. The B2B Institute (LinkedIn). IPA Databank B2B cases, 1998–2018; fewer than 50 cases. business.linkedin.com
  16. The B2B Institute (LinkedIn). “5 Principles of Growth in B2B Marketing”, summary page, which recommends a 50/50 split. business.linkedin.com
  17. Romaniuk, J. (2022). “Category Entry Points in a Business-to-Business (B2B) World.” Ehrenberg-Bass Institute, B2B report, July 2022, written with The B2B Institute. Survey of business insurance decision makers. marketingscience.info
  18. Microsoft Corporation (2026). Annual report on Form 10-K for the fiscal year ended 30 June 2026, Item 1, LinkedIn. US Securities and Exchange Commission. sec.gov
  19. Gordon, B. R., Moakler, R. and Zettelmeyer, F. (2023). “Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement.” Marketing Science, 42(4), 768–793. doi:10.1287/mksc.2022.1413
  20. LinkedIn Marketing Solutions. “Revenue Attribution Report.” Connects to HubSpot, Microsoft Dynamics 365 or Salesforce. business.linkedin.com
  21. LinkedIn Marketing Solutions. “LinkedIn Insight Tag.” Website conversion tracking, retargeting and visitor demographics. business.linkedin.com
  22. LinkedIn (2025). “Helping B2B Marketers Prove Campaign Performance with New Features.” LinkedIn Pressroom, February 18, 2025. Conversions API and optimizing for qualified leads. news.linkedin.com
  23. LinkedIn Marketing Solutions. “Testing”: A/B testing and Brand Lift testing. business.linkedin.com

More in Paid Media

Part of the Paid Media hub - see all 12 resources on this topic.