Ask the marketing team how many leads a campaign generated and you’ll get one number. Ask the sales team how many of those actually turned into anything and you’ll get a different number — sometimes a very different one. Neither team is lying. They’re usually looking at two systems that define “conversion” differently, and nobody has reconciled the two.
This is one of the most common — and most fixable — sources of friction between marketing and sales. Here’s where the disagreement actually comes from, and a practical way to close the gap.
Two systems, two definitions of “conversion”
A marketing platform typically counts a conversion the moment a form is submitted, an ad is clicked and a session completes a goal, or an email link is followed. That’s a real signal, but it’s an intent signal — it tells you someone raised their hand.
A CRM counts a conversion further down the line: a Lead becomes a qualified Opportunity, an Opportunity is marked Closed Won. That’s an outcome signal — it tells you money actually moved.
Both numbers are correct. They’re just answering different questions. The problem is when a report casually uses the word “conversion” for both and nobody notices they’ve been comparing intent to outcome the whole time.
UTM and source data: useful until it isn’t tracked consistently
UTM parameters — the tags on a URL that say which campaign, channel, and ad sent someone to your site — are the thread that’s supposed to connect a marketing touchpoint to a CRM record. They work well when every campaign tags its links consistently and that data is actually captured on the Lead or Contact record at the moment of conversion.
They stop working the moment someone shares a link without UTMs, a redirect strips the parameters, or the CRM has a “Lead Source” field but nobody built a rule for the marketing platform to populate it. At that point, the CRM shows a pile of Leads with source “Web” or blank, and marketing has no way to prove which campaign actually influenced the deal that closed six weeks later.
Leads, Contacts, and Accounts aren’t the same relationship the marketing platform assumes
Most marketing platforms think in terms of one record per email address. Salesforce, once a Lead converts, splits that person into a Contact tied to an Account — and a single Account might have five Contacts who each interacted with five different campaigns over a year. A report that tries to attribute a closed deal to “the campaign that converted the lead” is quietly ignoring the other four people who influenced the decision, because the data model on the marketing side was never built to represent that.
Lifecycle stages that don’t line up
Marketing automation tools usually have their own lifecycle stage field — Subscriber, Marketing Qualified Lead, Sales Qualified Lead — that’s meant to mirror Salesforce’s Lead Status and Opportunity Stage, but rarely stays in sync unless something is actively keeping them aligned. When the two drift, marketing dashboards show progress that sales dashboards don’t reflect, and vice versa, and both teams start distrusting the other’s numbers instead of noticing the sync problem underneath.
Duplicate records quietly break attribution
Someone fills out a form with a personal email, then later gets added to the CRM manually by a rep with their work email. Now there are two records, the marketing touches are attached to one and the closed deal is attached to the other, and the attribution report shows a “new” customer who supposedly never engaged with any marketing at all. Deduplication isn’t just a data-hygiene nicety here — it’s a direct input to whether attribution reporting means anything.
Offline conversions rarely make it back into the loop
A phone call, a trade show conversation, a referral that gets typed straight into Salesforce by a rep — none of that has a UTM tag, and none of it shows up in the marketing platform’s own reporting unless someone deliberately logs it there too. Over time, this systematically under-credits channels that generate high-value offline conversations (referrals, events, direct outreach) and over-credits anything with a clean digital trail, simply because the digital trail is easier to measure.
A worked example
Picture a mid-sized services firm running a paid search campaign. The marketing platform reports 40 form-fill “conversions” for the month. Sales looks at Salesforce and sees only 6 new Opportunities tagged with that campaign’s source, and concludes the campaign underperformed. In reality, several things are likely happening at once: some of the 40 were duplicate submissions from existing Contacts that never created new Leads, some correctly became Leads but haven’t been reviewed yet and their source field wasn’t set, and a few converted into real business but got logged under an Account a rep already had a relationship with, so the campaign’s UTM data never made it onto the record that eventually closed. The campaign might genuinely have underperformed — or it might have done reasonably well and be invisible because the two systems were never reconciled. Without doing the reconciliation work, there’s no way to tell which is true, and decisions about next month’s budget end up guessing.
A practical framework for reconciling the two
None of this requires a total rebuild. It requires agreeing on a few things once, in writing, and then enforcing them:
- Agree on what “conversion” means in each report. Marketing-qualified conversion and closed-revenue conversion are both legitimate metrics — the fix is labeling them clearly, not picking one and ignoring the other.
- Standardize UTM tagging and make sure it lands on the record. Every campaign link gets tagged the same way, and the CRM field it populates is defined once, not left to whatever the integration happened to do by default.
- Decide how multi-contact Accounts get attributed. Even a simple rule — first touch, last touch, or “any Contact on the Account counts” — is better than an undocumented default nobody agreed to.
- Map lifecycle stages explicitly, both directions. Marketing Qualified Lead should correspond to a specific Salesforce status, and that mapping should be the thing that updates automatically, not two separate manual processes.
- Run deduplication on a schedule, not just at import. New duplicates get created every week that new leads come in; dedup has to be ongoing, not a one-time cleanup.
- Give offline conversions a path back into the system. Even a simple “How did you hear about us?” field a rep fills in on a Lead does more for attribution accuracy than another dashboard.
Most of these are process and configuration fixes, not new software. The tools you already have — Salesforce, whatever marketing platform is connected to it — can usually do this correctly. What’s usually missing is someone treating the connection between them as a system to be designed, rather than a default that came out of the box.
Where this fits into a bigger picture
This is the same underlying problem behind our own digital marketing work — campaigns that report clean numbers but don’t reflect what actually closed. If your marketing dashboard and your Salesforce pipeline tell two different stories, it’s usually fixable without replacing either system. Tell us what’s not lining up and we’ll tell you honestly whether it’s a quick fix or something bigger.