Build a signal mix that fits your business

Funkel runs thirteen LinkedIn intent signal strategies. Most teams should run two or three of them at any given time. The question is which two or three. The answer depends on your sales-cycle length, your team shape, and how broad your ICP is. This playbook walks through the decision frame, then four worked recipes by category.

Step 1. Apply the three-axis decision frame

Three questions decide your mix. Answer them honestly before you turn anything on.

Cycle length. Fast-cycle products (close in weeks) reward signals that catch the buyer in transition:recent_job_changes, recently_raised_funds, your_profile_url. Slow-cycle products (close in quarters) reward signals that surface considered evaluation:competitor_pain, competitor_urls, influencer_profiles.

Team shape. Solo founders should run signals with a high bar to entry, low volume, and high relevance:your_profile_url, competitor_pain. SDR teams can run higher-volume signals because they have the inbox capacity to triage: recent_job_changes, company_followers.

ICP breadth. Narrow ICPs (one role, one vertical, one stage) work with content-discovery signals like keyword_tracking because the filter does the targeting work. Broad ICPs need intent signals to do the targeting; otherwise volume floods the funnel with wrong-fit candidates.

Your three answers tell you which two to three signals to pick. The four recipes below are starting points, not laws.

Step 2. Recipe for dev tools and engineering categories

Selling to engineering teams (devops, observability, platform, AI infra). Cycle is medium, ICP is broad-ish (any company doing engineering at scale), team is usually the founder for the first 100 customers.

The mix:

• competitor_pain as the primary. Engineering teams complain publicly about tools. Datadog, Sentry, PagerDuty, Snowflake; the names rotate every quarter. The five-query expansion catches the discontent without you needing to predict it.
• keyword_tracking as a baseline. Configure with the category vocabulary (“observability”, “platform engineering”, “SRE”, “incident response”). Catches people whose role is the right shape even when their title is not.
• recent_job_changes as a tiebreaker. New heads of platform are evaluating tools in their first quarter; the signal is loud and the buyer is in a deciding mindset.

Skip the engagement signals (company_followers, influencer_profiles) until you have a real content presence and recognizable category influencers. Without those, the signal produces noise.

Step 3. Recipe for vertical SaaS

Selling into one vertical (legal tech, healthcare ops, construction, hospitality). Cycle is medium-to-slow, ICP is narrow (specific role at specific vertical), team can be founder or first SDR.

The mix:

• recent_job_changes as the primary. New ops leaders at vertical companies inherit the stack; they are reviewing it in week one. The narrow ICP gates do the targeting work.
• influencer_profiles if your vertical has clear thought-leader gravity. Configure the influencer profiles whose audience is your buyer; the signal surfaces engaged readers, not the influencers themselves.
• competitor_urls if your competitor publishes regularly on LinkedIn. People engaging with their content have the budget and the problem; many are evaluating alternatives quietly.

Skip competitor_pain in tight verticals. Public complaints about a vertical-specific competitor are rare because the buyer pool is small and word travels.

Step 4. Recipe for agencies and professional services

Selling consulting, marketing services, design, product development. Cycle is medium, ICP is broad on industry but narrow on company stage (typically Series A through Series C), team is usually one or two people.

The mix:

• your_profile_url as the highest-intent signal. People who looked at your profile are evaluating you, not a category. The volume is low, the conversion is the highest of any signal.
• competitor_pain against the competing agencies. Buyers leaving an agency talk about it openly; the five-query expansion catches the frustration months before they actively shop.
• recently_raised_funds as the trigger for new budget. Newly funded teams are hiring services in the first six weeks of the raise. Lead with the work, not the money.

Skip volume signals entirely. Agency outbound at high volume reads as desperate; the signal mix above keeps weekly invites under forty per sender, which is exactly right for the category.

Step 5. Recipe for marketplaces and two-sided products

Selling a platform that needs both supply and demand (creator tools, marketplace SaaS, network products). Cycle is fast on supply, slow on demand, ICP is two distinct cohorts.

The mix:

• top_active_profiles for the supply side. Active content creators in your category are looking for distribution and tools; configure with the vocabulary they use about their craft.
• recently_raised_funds for the demand side. Newly funded teams have budget for new channels; the buyer is the head of marketing or head of growth in the first quarter post-raise.
• company_followers as the warm-side signal. People who engaged with your company posts already know who you are; the connection request lands warm regardless of which side they are on.

Run two agents (one per side) rather than mixing supply and demand signals in one campaign. The opener for a creator is not the opener for a buyer; do not force the AI to write across both.

What to drop after 14 days

Two weeks is the right window to evaluate a signal. Drop any signal that:

• Produces fewer than ten candidates per week. Either it is misconfigured or the underlying data is too thin.
• Has an acceptance rate ten points below your other signals. Different signals justify different openers; if you have not tuned the opener to the signal yet, do that before dropping. But a persistent gap means the signal is wrong for your category.
• Surfaces consistently wrong-fit profiles even after you tighten the ICP. Some signals are structurally wrong for some businesses; cut your losses.

Three concentrated signals beat ten diffuse ones. The teams that compound on Funkel are the ones that pick a tight mix, commit for two weeks, then iterate.

For a deeper read on what each individual signal predicts, see our field guide to LinkedIn intent signals.