Sales-Sourced Pipeline Is 62% of the Number. Marketing Is 19%.

The pipeline-sourcing split just moved, and most plans have not caught up

If you are building next year's pipeline model right now, one number should reframe the exercise before you touch a single coverage assumption: sales-sourced pipeline accounts for roughly 62% of total pipeline, against about 19% from marketing, with partner and other sources making up the remainder. That figure comes out of ICONIQ's State of Go-to-Market 2026 research, based on a survey of 155+ B2B SaaS executives spanning CROs, CEOs and RevOps leaders.

Look at the high-growth cut of the same data and the gap widens further. Companies posting top-quartile growth pull 60-80% of total pipeline from sales and channel-generated motions, versus 15-20% from marketing. Channel partnerships alone represent 27-31% of revenue for that group.

That is not a small attribution wobble. It is a structural statement about where the number comes from — and it lands at exactly the moment most GTM orgs are planning 2027 headcount, quota coverage and tooling spend against a mental model built when marketing-sourced pipeline was the headline metric on every board deck.

Here is the uncomfortable part. Most sales leaders will read "sales sources 62% of pipeline" and quietly assume that means their outbound engine is working. It usually does not mean that at all. It often means the burden of pipeline creation has shifted onto the sales org without the sourcing infrastructure shifting with it — and the metric is doing a poor job of telling anyone the difference.

This post breaks down what the 2026 sourcing data actually says, why the sourced-pipeline number is one of the least trustworthy figures in the revenue stack, and what a sales org should change this quarter to earn its 62% instead of merely being charged with it.

What the 2026 sourcing benchmarks actually show

The headline splits

Pull the credible benchmark sets together and a consistent picture forms, with real variance by motion:

Two things jump out.

First, motion drives the split far more than company quality does. A freemium, content-heavy business will report 60-75% marketing-sourced pipeline and be perfectly healthy. An enterprise business selling $200K+ ACV will report 15-30% and be equally healthy. Comparing yourself against a median that blends both is worse than useless — it is actively misleading.

Second, there is a crossing point. ZoomInfo's analysis notes that once average deal size crosses roughly $50K, marketing-sourced share drops from about 59% to about 47% as sales-led touch becomes necessary to create the opportunity at all. Every move upmarket mechanically shifts sourcing responsibility onto sales, whether or not anyone plans for it.

Who inside sales is actually doing the sourcing

"Sales-sourced" is not a single motion. Benchmark data on the internal split puts SDRs at roughly 40% of pipeline generation, marketing at 25-30%, and AEs self-sourcing the rest. That last bucket is where most of the quiet pain lives.

The economics of AE self-sourcing are brutal and well documented. AEs already spend only 28-30% of their week actually selling, per Salesforce's State of Sales research. Push self-prospecting onto them and you are buying prospecting hours at $36-54 per hour against roughly $14 fully loaded for an SDR doing the same work — and burning 8-12 hours per AE per week doing it. Ratio guidance across 2026 benchmark sets lands around 1 SDR per 2.4-2.6 AEs on average, tightening to 1:1.5-1:2 for SMB outbound and loosening to 1:3-1:4 for enterprise inbound-heavy motions.

So the 62% figure frequently decomposes into something like: a functioning SDR layer doing 40%, a marketing layer doing less than it used to, and a set of AEs absorbing the delta at the worst possible cost per hour. That is not a GTM strategy. That is a gap being papered over with expensive labour.

Why sourced pipeline is a metric that lies

Before anyone reorganises around the 62/19 split, it is worth saying plainly: sourced pipeline is an artifact of your attribution configuration, not a fact about your business.

Change what counts as a "touch," change which system logs first, or change the lookback window, and the same company's sourced percentage swings by 20 points without a single thing changing in the real world. Three common distortions:

• Last-touch-before-opp wins by default. An SDR books a meeting with a prospect who found you through a webinar, three podcast appearances and two months of newsletter reading. The opportunity is tagged sales-sourced. Nothing about that tag is wrong, and nothing about it is informative.
• Attribution windows quietly decide the answer. A 30-day lookback and a 180-day lookback produce materially different pictures of the same funnel, particularly in long enterprise cycles.
• Self-serve and PLG entry points get orphaned. Product-qualified accounts that a rep later expands frequently land in whichever bucket the CRM admin last configured, which is rarely a deliberate choice.

This matters practically, not just philosophically. If you use a distorted sourcing number to reallocate budget from marketing to sales headcount, you can starve the demand-creation layer that was quietly making your outbound work — and then watch reply rates fall while you add reps.

The right way to hold both ideas at once: treat the 62/19 split as directional evidence that pipeline creation responsibility has moved toward sales, and treat your own internal sourced number as a management artifact that needs auditing before it drives a decision.

A three-question audit before you act on your own number

1. What is our attribution model and lookback window, in one sentence, and when did we last change it?
2. If we re-ran last four quarters of closed-won with a marketing-influenced model instead of marketing-sourced, how much does the picture change?
3. Of our "sales-sourced" opportunities, what percentage involved a prospect who had a prior interaction with our content, community or product?

If the answer to question three is above 40%, your sales team is not sourcing cold pipeline. It is converting warm demand — which is good, and which you should build deliberately rather than by accident.

The real implication: sales owns sourcing, but not the sourcing infrastructure

Here is the structural problem the benchmark data exposes. Over the past decade, demand generation built serious infrastructure: attribution systems, intent monitoring, nurture sequencing, scoring models, content engines. Sales sourcing got a contact database and a sequencer.

Now sourcing responsibility has moved to sales, and the infrastructure did not move with it. That is why so many teams read "62% sales-sourced" as a success signal when it is frequently a load-bearing crack.

What sales-side sourcing infrastructure actually needs to cover:

• Detection. Something has to notice that an account moved — hiring for a role that implies your problem, complaining about a competitor, engaging with a relevant post, announcing a funding round, reorganising a team.
• Qualification against ICP. A signal is not a lead. Most detected activity is noise until it is scored against who actually buys from you.
• Research. The difference between a 2% reply rate and a 12% one is almost never the sequence template. It is whether the first line demonstrates you understand this specific account's situation.
• Timing. Signals decay. A hiring post is worth far more in week one than week six, and most teams act on week six because that is when the weekly list refresh runs.
• Safe execution. Particularly on LinkedIn, where enforcement through 2026 has been behavioural rather than volumetric — burst patterns, data-centre IPs and high ignored-invitation ratios draw restrictions more reliably than raw volume does.

Stacking point solutions across those five layers is exactly how teams end up paying for Sales Navigator, an enrichment platform, a research tool, a sequencer and a scraper, then hiring someone to keep the joins from breaking. Platforms like Updately exist to collapse that chain — capturing intent signals across LinkedIn, Reddit and X, scoring them against ICP, researching the prospect before writing, and sending inside platform limits — but the point stands regardless of what you buy. If sales owns 62% of pipeline creation, sales needs a sourcing system, not a contact list and a hope.

The playbook: earning your 62% instead of absorbing it

1. Segment your sourced-pipeline target by motion, not by company average

Stop benchmarking against a blended median. Set the sales-sourced target by segment and ACV band. An enterprise team above $200K ACV should expect to self-source 70-85%. A mid-market team in a content-strong category might reasonably sit near 50%. Holding both to the same number guarantees one of them is being managed against fiction.

2. Fix the SDR-to-AE ratio before you fix the sequences

If your AEs are spending more than four hours a week on self-prospecting, the ratio is wrong and no amount of messaging coaching will compensate. Run the arithmetic explicitly: AE hours spent prospecting × fully loaded hourly cost, against the cost of tightening the ratio. In most mid-market orgs the ratio fix pays for itself before the messaging work even starts.

3. Move from list-sourced to signal-sourced

The single highest-leverage change available to most sales orgs this quarter. Instead of starting from a filtered list of accounts that match a firmographic profile, start from accounts that did something. The difference in reply rate is not marginal — it is the difference between interrupting someone and arriving at a moment when the problem is live.

Practical signal set to run first, in rough order of conversion quality:

• Prospects who viewed your profile or engaged with your content in the last seven days
• Accounts posting publicly about the pain you solve, on LinkedIn, Reddit or X
• Accounts mentioning or complaining about a competitor
• Hiring signals that imply your problem is now funded and owned
• Job changes into buyer-adjacent roles at accounts already in your ICP
• Funding or reorganisation events that unlock budget

4. Instrument signal-to-touch latency as a first-class metric

Most teams measure activity volume and reply rate. Almost none measure how long it takes to act on a signal once it fires. That latency number predicts outbound performance better than volume does. Set a target — 24 hours for high-intent engagement signals, 72 hours for hiring and funding events — and report against it weekly.

5. Build the marketing-to-sales handoff you would want if sourcing were reversed

Since a meaningful share of "sales-sourced" opportunities involve prior marketing exposure, make that exposure visible to the rep at the moment of outreach. A rep who knows the prospect read three specific pieces of content writes a different first message than one who does not. This is cheap to build and disproportionately effective.

6. Audit channel before you assume it is small

For high-growth companies, channel represents 27-31% of revenue. If your partner motion contributes under 10% of pipeline and you sell to mid-market or enterprise, that is a gap worth a dedicated plan rather than a line item.

Takeaways for the next 30 days

• The split is real and directional. Sales-sourced pipeline at ~62% against marketing at ~19% reflects a genuine shift in who creates pipeline, not a measurement quirk. Plan for it.
• Your own sourced number is less reliable than you think. Audit the attribution model and lookback window before any budget or headcount decision rests on it.
• Benchmark by motion, never by blended median. A 25% marketing-sourced share is excellent for enterprise and alarming for PLG.
• AE self-sourcing is the most expensive pipeline you buy. Check the SDR-to-AE ratio first; it is usually the cheapest fix available.
• Sourcing responsibility moved to sales; infrastructure did not. Detection, ICP scoring, research, timing and safe execution all need to exist somewhere in your stack, whether you build or buy them.
• Signal-sourced beats list-sourced on every metric that matters. Start with profile viewers, content engagers, competitor mentions and hiring signals — the four highest-intent, lowest-effort signals available to almost any B2B team.

The teams that will outperform against this data are not the ones that read "62%" and add SDRs. They are the ones that read it as a mandate to build a real sourcing system on the sales side — one that finds the moment, understands the account, and reaches out while the problem is still live.

That is the whole difference between owning pipeline creation and merely being blamed for it.

Sources: ICONIQ State of Go-to-Market 2026 · ICONIQ Building the Modern GTM Org · ZoomInfo marketing-sourced pipeline analysis · Prooflytics 2026 marketing-sourced benchmarks · ModernLeads SDR-to-AE ratio benchmarks 2026