The number on our homepage, taken apart

How ×3.9 adds up, lever by lever

The DemandSense pipeline model is built on averages from independent studies and industry benchmarks.

Pipeline a month over six months, same ad budget: ×1.0, ×1.5, ×2.1, ×3.1, ×3.9, ×3.9. Split into four levers: Fewer wasted impressions ×1.18, Your own audiences ×1.60, LinkedIn learns from your deals ×1.64, You move budget to what works ×1.25 (our assumption).Month 1Month 6×3.9pipeline a monthsame budget×1.18×1.60×1.64×1.25our assumption
  • ×1.25 You move budget to what works (our assumption)
  • ×1.64 LinkedIn learns from your deals
  • ×1.60 Your own audiences
  • ×1.18 Fewer wasted impressions

Month by month

What happens, and when

Pipeline a month over the model's six months, same ad budget: ×1.0, ×1.5, ×2.1, ×3.1, ×3.9, ×3.9, built up lever by lever.Month 1Month 2Month 3Month 4Month 5Month 6×1×2×3×4×1.0×1.18 Fewer wasted impressions×1.60 Your own audiences×1.64 LinkedIn learns from your deals×1.25 You move budget to what worksour assumption

  1. Week 11 / 6

    Pipeline hasn't moved yet

    Ads stop running to nobody

    Ads stop running at night, on weekends and to companies outside your ICP, and one company can't take most of the impressions. Pipeline follows about a month later.

    ×1.18 · Fewer wasted impressions

    In open-web programmatic, 35% of every dollar that enters a DSP goes to non-viewable, invalid, unmeasurable or made-for-advertising traffic.

    Independent + our testsLinkedIn ads scheduling →

    ×1.60 · Your own audiences

    Custom audiences are the key element of account-based marketing (ABM). ABM programs lift pipeline by over 11% (Gartner). Henkel's ABM program got a lead-form click-through rate 1.45× the industry benchmark.

    Analysts + judged caseswebsite visitor identification →

  2. Month 22 / 6

    ×1.5 pipeline a month

    The first test reads

    Week 5: the first test reads on engaged companies, long before any deal. You start moving budget to the campaign that works.

    ×1.25 · You move budget to what works

    A test reads in about 5 weeks on engaged companies, instead of waiting for deals.

    AssumptionLinkedIn ads analytics →

    In this month's pipeline

    • Fewer wasted impressions×1.18
    • Your own audiences×1.30
    • LinkedIn learns from your deals×1.00
    • You move budget to what works×1.00

  3. Month 33 / 6

    ×2.1 pipeline a month

    LinkedIn starts learning from your deals

    Your closed deals start flowing back to LinkedIn, so it learns who buys, not who clicks.

    ×1.64 · LinkedIn learns from your deals

    −39% cost per qualified lead, so the same budget buys 1.64× the qualified leads.

    Randomized study + LinkedInLinkedIn Conversions API →

    In this month's pipeline

    • Fewer wasted impressions×1.18
    • Your own audiences×1.60
    • LinkedIn learns from your deals×1.00
    • You move budget to what works×1.13

  4. Month 44 / 6

    ×3.1 pipeline a month

    Budget sits with what works

    Your budget now sits with what works. LinkedIn is halfway through learning from your deals.

    In this month's pipeline

    • Fewer wasted impressions×1.18
    • Your own audiences×1.60
    • LinkedIn learns from your deals×1.32
    • You move budget to what works×1.25

  5. Month 55 / 6

    ×3.9 pipeline a month

    All four levers at full strength

    The four levers multiply: 1.18 × 1.60 × 1.64 × 1.25. From here the monthly lift stops growing, and it stays on the same ad budget.

    In this month's pipeline

    • Fewer wasted impressions×1.18
    • Your own audiences×1.60
    • LinkedIn learns from your deals×1.64
    • You move budget to what works×1.25

  6. Month 66 / 6

    2.9× the pipeline, added up

    It holds

    The monthly multiple holds. Added up since month 1, that's 2.9× the pipeline you'd have had by now.

    In this month's pipeline

    • Fewer wasted impressions×1.18
    • Your own audiences×1.60
    • LinkedIn learns from your deals×1.64
    • You move budget to what works×1.25
  • Marketixa
  • Aussie Founders

Where's the catch?

Without our assumption, it's ×3.1

«You move budget to what works» is the one lever no published study sizes. Take it out and the sourced levers give ×3.1 pipeline a month. Budget caps protect your spend and add nothing to the multiple.

Three cases, same four levers

The model has a range. The homepage shows the optimistic case; here is what the others assume, lever by lever.

×1.4

×2.3

×3.9

Conservative

Middle

Optimistic

  • Fewer wasted impressions

    ×1.05

    ×1.11

    ×1.18

  • Your own audiences

    ×1.10

    ×1.45

    ×1.60

  • LinkedIn learns from your deals

    ×1.18

    ×1.31

    ×1.64

  • You move budget to what works · our assumption

    ×1.00

    ×1.10

    ×1.25

  • Sourced levers only

    ×1.4

    ×2.1

    ×3.1

How each lever saves

What changes in your ads, the research behind the number, and the DemandSense tool that does it.

  • SMTWTFSNightDayNightAds runAds off

    Fewer wasted impressions×1.05–×1.18

    How it saves

    Ads stop running at night, on weekends and to companies outside your ICP, and one company can't take most of the impressions.

    The research

    • ANAANA, Programmatic Transparency Study, 202335% of every dollar that enters a DSP goes to non-viewable, invalid, unmeasurable or made-for-advertising traffic. 21 advertisers, $123M.
    • DemandSenseDemandSense, our own testsFour two-week scheduling tests: 44–67% more impressions on the same budget.
    LinkedIn ads scheduling →
  • Broad job-title audienceYour own audiencesCompanies already in playEveryone else

    Your own audiences×1.10–×1.60

    How it saves

    Companies that visit your site and the accounts already in your pipeline go to LinkedIn as your own audiences, so the budget lands on companies already in play instead of a broad job-title audience.

    The research

    website visitor identification →
  • Your CRMclosed dealsLinkedInClosed deals, people like themEveryone else

    LinkedIn learns from your deals×1.18–×1.64

    How it saves

    Your closed deals flow back to LinkedIn, so it learns who buys, not who clicks, and spends the same budget on people like your buyers.

    The research

    LinkedIn Conversions API →
  • Test readsabout 5 weeksDeals closeCampaign ACampaign Bmonths earlier

    You move budget to what works×1.00–×1.25

    How it saves

    A test reads in about 5 weeks on engaged companies instead of waiting for deals, so you move budget to the campaign that works months earlier.

    Why it works

    The more experiments you run, the sooner you find what works. Waiting for deals, one test takes a whole sales cycle. Read on engaged companies, it takes about 5 weeks, so in the same time you run several tests instead of one, and each one moves budget to the better campaign.

    LinkedIn ads analytics →

There's more

What the multiple leaves out

These help your pipeline too, but no published number sizes them honestly, so none of them is in ×3.9.

What a skeptic asks

No. ×3.9 is the optimistic case of a model built from independent research: a randomized study of 70,909 experiments on live ad campaigns, an industry audit of $123M in ad spend, analyst benchmarks and judged cases for account-based programs, and LinkedIn's own figure for one lever. Our own tests measured delivery, not pipeline: four two-week scheduling tests showed 44–67% more impressions on the same budget. Your account will have its own numbers.

The model uses averages. Your account has its own numbers.

Connect your LinkedIn ads and see which target companies your campaigns reach this week.

Start free

30-day trial · No credit card · All features