Why Don’t Marketing and Sales Pipeline Numbers Match? How to Fix It

Your pipeline (the open sales opportunities and their value) has two owners and two numbers. Marketing brings $2.4M to the quarterly review. Sales brings $1.1M. Both pulled their number from the same CRM that morning, and the first 20 minutes of the meeting go on arguing about which one is real, while the decisions the meeting was called for wait.

The fight is costly, and it isn’t only about time. In Validity’s 2026 survey of 500 marketers, 38% said campaign data is at times manipulated to make results look better to leadership, rising to 67% among C-suite respondents, and nearly a third of teams spend six or more hours a week just fixing and reconciling data. Underneath it sits a definitions problem: in Gartner’s May 2025 survey of 243 sales leaders, 49% of CSOs said their sales team’s definition of a qualified lead differs greatly from marketing’s. If the two teams don’t agree what “qualified” means, they won’t agree what “pipeline” means either.

The good news is that the gap is almost always explainable, line by line, with data you already have. This guide shows the five places the two reports count differently, how to settle sourced versus influenced pipeline, which date and which source field to use, and a 30-day plan to get to one number. It includes a worked sample bridge you can copy.

What are the signs your pipeline numbers don’t match, and where does the gap come from?

The two reports differ in five places: the object they count, the credit rule, the date, the stages included and the source field. Find which of the five apply and you can explain the whole gap, usually without finding a single wrong number.

Most pipeline arguments start from the assumption that one team has bad data. Sometimes that’s true. Far more often, both reports are accurate answers to different questions. Marketing’s report asks “how much pipeline did marketing touch?” and counts from the moment a deal is created. Sales’ report asks “how much qualified pipeline do we have to close?” and counts from the moment a rep accepts it. Both are reasonable. They just aren’t the same thing, and they’ll never match while they share a name.

Pull both reports for the same quarter, then go down the table and find every row that matches what you see. Most teams find two or three at once.

The thresholds are practitioner rules of thumb, not industry standards. They tell you where to look first; they aren’t pass-or-fail scores.

Object and credit signs

The biggest gaps usually come from the first two rows. Some marketing dashboards still show “pipeline” built from contacts or MQLs multiplied by an average deal size. That’s a forecast of possible pipeline, not pipeline, and it will always run ahead of what sales sees. The second cause is credit. A marketing report that counts every deal marketing touched will be far larger than a sales report that credits marketing only when it created the first touch. Neither is wrong, but adding influenced pipeline to the company total double-counts every deal that sales also sourced.

Date and stage signs

If the totals differ but the deal lists mostly overlap, look at dates and stages. A deal created on the last day of March lands in Q1 on marketing’s report and in Q2 on sales’ report if it was qualified in April. A report that includes every open deal from the discovery stage will be bigger than one that starts at qualified, and the difference swells at quarter-end, when reps create placeholder deals to hold a meeting.

Field signs

The quietest cause is the source field itself. Many CRMs have three or four fields that each claim to hold “the source”, and reps can often edit the one on the deal. If deals regularly change from a marketing source to “Outbound” or “Referral” after a rep takes them, both teams will be reporting honestly from different fields, and the gap will keep moving every week.

The cost of leaving this alone shows up well beyond the meeting room.

Sourced vs influenced pipeline: which number is right?

Both are right, because they answer different questions. Sourced asks “who started this deal?” Influenced asks “did marketing play any part in it?” Report them side by side, and never add them together.

This is where most of the gap in most companies lives, so it’s worth defining both terms precisely.

Marketing-sourced pipeline counts deals whose first known source was a marketing channel: organic search, paid ads, an event, a webinar, a content download. Every deal has exactly one source, so sourced numbers across teams add up to 100% of pipeline.

Marketing-influenced pipeline counts every deal where marketing touched any contact on the deal within a set window, usually something like the 90 days before the deal was created, or any time before it closed. A deal sourced by an SDR’s cold email is still “influenced” if the buyer later attended a webinar. Influenced numbers overlap with every other team’s, so they never add up to anything.

Why influenced pipeline runs so high

B2B buying involves many people and many touches. Dreamdata’s 2026 benchmarks, drawn from 3.5 million customer journeys in its own customer data, count 88 touchpoints across 4 channels and 10 stakeholders in the typical B2B journey, and find that 81% of that journey happens before a deal enters the sales pipeline. With that many touches spread across that many people, almost every deal has a marketing touch somewhere. An influenced number that covers 60–70% of all pipeline isn’t inflated; it’s what modern buying looks like.

Why sourced pipeline runs low

Sourced credit goes to one first touch, usually on one contact, while ten people buy. If the first person from an account found you through a peer recommendation and the next nine read your content for months, the deal is “referral sourced”. Software can also miss the first touch entirely when it happens in a podcast, a private community or an AI chatbot. That’s why sourced numbers tend to undercount marketing’s role, and why marketing teams reach for influence.

What each number is for

Use sourced pipeline for targets and budget: it’s the number marketing can own, it adds up with sales’ and partners’ numbers, and it doesn’t double-count. Use influenced pipeline to see which programs show up in deals that win, and which never appear. Neither proves cause. A webinar that appears in 40% of won deals may have helped close them, or may just attract buyers who were already going to buy.

Why do created, qualified and closed dates give different totals?

Each team picks its own date, so one deal lands in three different quarters. Pick one date per metric, write it down, and use it everywhere.

Every deal carries at least three dates that matter for reporting, and each team has a natural favourite.

• Create date. When the deal record was made. Marketing likes it because it’s the earliest moment a deal exists, and it’s closest to the campaign that produced it.
• Date entered the qualified stage. When sales agreed the deal is real. Sales likes it because it excludes deals that never should have been created.
• Close date. When the deal was won or lost. Finance uses it, because that’s when revenue is booked.

Take one deal created on 18 February, qualified on 9 April and won on 27 August. Marketing’s “pipeline created” report counts it in Q1. Sales’ “qualified pipeline” report counts it in Q2. Finance books the revenue in Q3. Multiply that by a few hundred deals with sales cycles of several months, and the quarterly totals drift apart even when every deal and every amount is identical.

The rule that ends it

For pipeline, use the date a deal enters the qualified stage. It’s late enough to exclude placeholder deals and early enough to give marketing credit in the quarter the work turned into a real opportunity. For bookings and revenue, use the close date. Keep create date for operational reports, such as how fast new deals get qualified, and keep it out of the leadership number. HubSpot stores a “Date entered” property for every deal stage, and Salesforce can track the same with stage history, so you don’t need anything new to report this way.

Snapshots versus period totals

There’s a second date trap. “Open pipeline today” is a snapshot: the value of all open qualified deals right now. “Pipeline created this quarter” is a period total: the value of deals that became qualified between two dates. They are different metrics, and a snapshot will move every time a deal closes, slips or changes amount. Label each report as one or the other.

Which source field should we report on?

Report on one source field per deal: a copy of the primary contact’s original source, written onto the deal when it’s created and then locked. Show the buyer’s self-reported answer beside it.

Most CRMs carry several fields that look like “source”, and each tells a different story about the same deal. In HubSpot, the usual four are below; Salesforce has close equivalents.

• Original source (contact). Set automatically by tracking on the contact’s first known visit or conversion. Stable, and the right starting point.
• Latest source (contact). Updated with each new session. Useful for nurture, wrong for pipeline credit, because it changes every time the buyer clicks an ad.
• Lead source (deal). Often a picklist that reps fill in or edit. It drifts toward whatever the rep remembers, which is usually their own outreach.
• “How did you hear about us?” (form). The buyer’s own answer. It catches sources software can’t see: podcasts, peers, communities and AI chatbots.

How to set up one locked field

Create a deal property, for example “Pipeline source”. Use a workflow to copy the primary contact’s original source into it when the deal is created. Make it read-only for reps. Let RevOps correct it, with a required reason logged on the record. From then on, every pipeline report uses that field and only that field. If a deal was really sourced by an SDR, the SDR’s own touch should show up as the contact’s first activity; if it doesn’t, fix the logging, not the report.

Why the self-reported field matters

Dreamdata’s finding that 81% of the journey happens before the pipeline is the reason software attribution misses so much. A required, free-text “How did you hear about us?” field on demo and trial forms is the cheapest way to see that hidden part. Don’t use it to replace the tracked source. Show it next to the tracked source, and look for patterns: when 20% of buyers say “a podcast” and your tracked data shows zero podcast-sourced deals, that’s an insight neither field gives alone.

Clean data is part of the fix

None of this works if the underlying records are messy. In Validity’s 2026 survey, 62% of organizations reported losing revenue directly due to poor CRM data quality, and only 41% have a dedicated data governance team or owner specifically to manage regulatory and privacy risk. If nobody owns the data, nobody owns the definitions built on it.

How do we reconcile the two pipeline reports?

Export both reports for the same quarter at deal level, match them deal by deal, and give every part of the gap a named reason and an amount. Anything left over is a real data error to fix.

A bridge turns an argument into a table. Instead of debating which total is right, you show exactly how one total becomes the other. It takes an afternoon with a spreadsheet the first time, and much less after that.

Step 1: Export both reports at deal level

For the same period, export every deal in each report with deal ID, amount, stage, create date, date entered the qualified stage, close date, deal type and every source field. Totals alone can’t be reconciled; deal lists can.

Step 2: Match on deal ID

Put both lists side by side. You’ll get three groups: deals in both reports, deals only in marketing’s and deals only in sales’. For deals in both, check whether the amounts match; amount changes between report dates are a common hidden gap.

Step 3: Tag each difference with one of the five gaps

For every deal in one report only, ask why. Influenced but not sourced? Not yet qualified? A renewal? Created in a different quarter? Source changed by a rep? Each difference should fit one of the five gaps from the first section.

Step 4: Sum each gap

Add up the amounts by gap. You now have a bridge: marketing’s total, minus each named gap, equals sales’ total. If something remains that fits no gap, it’s a genuine data error: a duplicate deal, a wrong amount or a missing association. Fix it at the source.

Step 5: Write the definitions and rebuild one report

Decide, gap by gap, which rule the company will use. Then write the definitions down and build one report on them. The bridge you just made becomes the evidence for each decision.

Worked example: a sample pipeline bridge

Here’s what that looks like for a fictional Series B company, “Vantlow,” selling field-service scheduling software and running its CRM on HubSpot. The numbers are illustrative sample data, built to show the method. Marketing reported $2.40M of Q3 pipeline. Sales reported $1.10M of marketing-sourced pipeline for the same quarter.

The striking thing about Vantlow’s bridge is that none of the $1.30M gap was a wrong number. $0.75M was a different question: deals marketing influenced but didn’t start. $0.43M came from counting rules, discovery-stage deals and renewals, that one report included and the other didn’t. And $0.12M came from a field reps could overwrite, where deals that began with a marketing touch had been re-labelled “Outbound” after the first call.

Once the gaps were named, the decisions were quick. Leadership now sees marketing-sourced qualified new-business pipeline of $1.22M as one line, and marketing-influenced pipeline of $0.75M as a separate line underneath, never summed. Sales accepted the higher sourced number because the locked field showed the first touch on every deal; marketing accepted the lower total because influence was still visible, just not counted twice. The next quarterly review started with decisions instead of a debate.

Which report should leadership use, and who owns the definitions?

Leadership should use one deal-based report with one written definition per metric. RevOps owns the definitions; the heads of marketing and sales sign them off together.

The one report

Keep it short. A leadership pipeline report needs five things:

• Qualified new-business pipeline created, counted on the date each deal entered the qualified stage, for the month and the quarter.
• Split by pipeline source: marketing, sales, partner, and any other source you track, adding up to 100%.
• Marketing-influenced pipeline as a separate line, clearly labelled, never added to the total.
• Win rate and average deal size by source, so a source that creates lots of pipeline but rarely wins is visible.
• Self-reported source summarised next to the tracked source.

Everything else, including lead counts, MQLs and channel-level campaign data, belongs in team reports, not the leadership one.

The metric dictionary

Write the definitions in a single shared document. Six to eight metrics are enough for most Series A–B companies. For each one, record the name, a one-sentence definition, the object it counts (deals), the date field, the stage filter, the source field, the owner and where the report lives. When someone asks “what does pipeline mean here?”, the answer is a link, not a meeting.

Who owns it

RevOps writes and maintains the dictionary and builds the report, because RevOps sits between the teams and owns the CRM. The head of marketing and the head of sales sign it off together, because a definition only one of them agreed to will be disputed in the first bad quarter. Finance checks that closed-won bookings in the CRM match booked revenue. As a rule of thumb, change a definition only at the start of a quarter, and note the change and the date in the dictionary so trends stay comparable.

That shared ownership is what most teams are missing today. Gartner’s 2025 survey found 49% of CSOs say their sales team’s definition of a qualified lead differs greatly from marketing’s. And alignment looks very different depending on who you ask: in Databox’s 2025 survey, only 38% of sales leaders said their teams are highly aligned on metrics, against 61% of RevOps leaders.

Putting it to work: 30 days to one pipeline number

Four weeks, one owner, definitions before dashboards. By the end, the company has a signed metric dictionary, a locked source field and one report leadership reads.

You can run this plan with one owner, usually the head of RevOps or marketing operations, plus an hour a week each from the heads of marketing and sales. You don’t need a new tool. Everything below works in the CRM you already have, with a spreadsheet alongside; it’s the groundwork any change to your CRM and reporting systems should start from.

Week 1: Put both reports side by side

Export marketing’s and sales’ pipeline reports for last quarter at deal level. Match them on deal ID and build the bridge: tag every difference with one of the five gaps and sum each one. Share the bridge with both teams before anyone proposes a fix.

Week 2: Write the metric dictionary

Agree six to eight metrics, starting with qualified pipeline created, marketing-sourced pipeline, marketing-influenced pipeline, bookings and win rate. For each, write the definition, object, date, stage filter, source field and owner. Get the heads of marketing and sales to sign it off.

Week 3: Lock the source field

Create the “Pipeline source” deal property, build the workflow that fills it from the primary contact’s original source on creation, and make it read-only for reps. Add a required “How did you hear about us?” field to demo and trial forms. Set up the monthly pipeline snapshot.

Week 4: Ship one report

Build the leadership report from the dictionary. Re-run last quarter’s bridge with the new definitions to show the before and after. Retire the old dashboards, or rename them so they can’t be mistaken for the leadership number.

The checklist

• Both pipeline reports exported at deal level for the same quarter
• Bridge built: every part of the gap named and summed
• Sourced and influenced pipeline reported as separate lines
• One date per metric; pipeline counts on entry to the qualified stage
• “Pipeline source” deal field filled on creation and locked
• Required “How did you hear about us?” field on demo and trial forms
• Metric dictionary signed off by the heads of marketing and sales
• Monthly pipeline snapshot stored; old dashboards retired

If you’d rather have someone build the bridge with you, that’s exactly what the free roadmap call is for.

Sources

• Validity, The State of CRM Data Management in 2026, press release (25 Aug 2026) — 500 B2B and B2C marketing professionals in the US, UK, Brazil, Australia and New Zealand
• Databox, Beyond Attribution: The Disappearing Buyer Trail (29 Oct 2025) — 300+ GTM leaders
• Dreamdata, LinkedIn Ads Benchmarks Report 2026, press release (10 Mar 2026) — 66M+ sessions, 3.5M+ customer journeys
• Gartner, CSO survey press release (21 May 2025) — 243 CSOs and senior sales leaders