I talked to Nick Turner, CEO at Dreamdata, about a fight that shows up in almost every B2B company I work with: marketing sourced vs. sales sourced. Teams argue over who gets credit for a deal because budgets, headcount, and paychecks depend on the answer.
Nick's position is blunter than most. When someone says "I sourced it," he thinks they're usually wrong, whichever team they sit on.
He spent years in sales leadership before running Dreamdata, which gives his view some weight on both sides of the argument.
We covered why the sourcing debate exists, how Dreamdata rebuilt its SDR motion around it, what compensation does to behavior, and how to measure performance when no single attribution model tells the full story.
"Who sourced it?"
Nick's first answer to the sourcing question was that neither label holds up. Pure inbound does happen.
Someone lands on the website without ever hearing from an SDR, books a demo through a scheduling link, and sales gets the meeting handed over.
Pure cold outbound can happen too, but Nick says it isn't effective.
If a company has a clear ICP, runs campaigns at it, and keeps its sales team disciplined about prospecting only into that ICP, very few accounts reach sales without some marketing contact first. In his view, sourcing fights usually point to a poorly defined ICP or a sales team working outside it.
That matches what I see in accounts. When sales and marketing share one target account list, marketing has often been running ads, events, and content at those accounts for months before an SDR sends the first email.
The CRM still says "sales sourced" because that's where the email got logged.
Nick made the same point about the CRM. It's a sales tool, built to track sales activity, and the sourcing field is one dropdown away from being changed.
"Everyone needs to accept that it's a relay race," he said. He'd rather use the data to decide where to invest next: which campaign, which channel, which ICP segment deserves more money.
Front-running the demo request
One of Dreamdata's biggest bets this year was building an SDR team.
Nick's reasoning started with timing. By the time a buyer requests a demo, they've already formed a view of the market, and Dreamdata wanted a chance to shape that view earlier.
So he set an expectation with the marketing team up front. "I expect the SDRs to front-run the demo requests, and that's by design," he told them.
SDRs now reach out when an account hits a set engagement score or takes specific actions that Dreamdata's own data links to closed-won deals.
He also told the company what the numbers would do. Inbound demo requests would fall and SDR-sourced meetings would rise.
Both happened, and total pipeline volume increased dramatically, because SDRs were catching buyers who had real interest but weren't ready to book a demo on their own.
What I like here is that Nick named the trade-off before it showed up in a report.
A last-touch dashboard would show inbound demos dropping and flag it as a marketing problem. Dreamdata decided in advance that the drop was part of the plan.
When self-sourcing pays more
I shared a story from my own sales days. Before I moved into marketing, I worked as an SDR, then an AE, then in sales management.
At one company, AEs earned a higher commission percentage on deals they sourced themselves, meant to keep them from waiting on SDRs.
It created a credit fight inside sales. I'd be working an account on LinkedIn for months, an SDR would get a reply to a cold email, and we'd argue over whose deal it was by pointing at CRM records.
AEs started avoiding the SDR team, which was the opposite of what the plan intended.
Nick called that a poorly designed comp plan, and he offered a simple test for anyone considering self-sourcing incentives: look at your conversion rates.
Compare deals that came in cold against deals where marketing and brand were already involved. In his experience, the supported deals convert at a higher rate.
He also flagged a cost that shows up later. Rewarding self-sourcing can push reps toward accounts outside the ICP, and the business feels that in retention.
Nick questioned whether AEs should be prospecting at all when SDRs and marketing are generating pipeline for them, with an exception for early-stage companies where a founding AE runs the whole cycle.
He admitted plenty of sales leaders would roll their eyes at that. I'm not fully sold either from my days in sales.
Some AEs build relationships that no SDR sequence would, and I'm not sure a comp plan can tell those apart from reps chasing non-ICP logos.
The 272-day problem with revenue-based comp
Most marketing leaders I talk to are measured on something different from sales: MQLs, SQLs, pipeline, or some mix. I asked Nick whether that split makes sense.
He answered through SDR comp first. Nick doesn't like paying SDRs on closed-won revenue.
Dreamdata's sales cycle runs about 80 days, and by the time a deal closes, the SDR who booked the meeting has made hundreds of dials since.
Dreamdata pays SDRs monthly, and he noted some companies pay weekly to keep motivation up in a hard job.
Marketing faces the same timing gap on a longer clock. According to Dreamdata's data, the average B2B customer journey runs about 272 days from first touch.
In Nick's view, tying a marketer's bonus to revenue that closes nine months after a campaign ran creates weak incentives, though marketers should still track what happens downstream.
At the executive level, Dreamdata takes a different approach. Every executive, including the heads of product and engineering, shares the same goals: ARR, retention, and a few others.
If the company hits them, every executive can hit their bonus.
Below that level, a demand gen team's job is qualified pipeline from the right ICP, with an eye on revenue but no quota tied to it.
Which attribution model should you use?
Nick's short answer: more than one. Dreamdata lets customers run three, four, or five attribution models at a time and compare where they agree.
He considers it a mistake to make budget decisions from a single model.
For B2B teams, no single attribution model gives a full picture of marketing performance. The more reliable approach is to triangulate across several models (first touch, last touch, linear, and multi-touch) and act where they point the same direction.
He was also critical of his own industry. Some martech vendors, he said, use terms like incrementality loosely, labeling tools as incrementality measurement when they aren't.
After 20 years in martech, Nick traces many of the measurement problems marketers face today back to promises vendors made and couldn't keep.
On marketing mix modeling, Nick was clear that he isn't an MMM expert. His rough threshold is around $1 million in spend before MMM becomes useful, which puts it out of reach for many B2B teams.
Long journeys with many people on the buying committee make it harder still, and MMM can't get down to campaign-level decisions. Where it works, it can help weight your attribution models by channel.
I've lived the cost of relying on one model. At one company, the board had aligned on last touch, with goals built on MQLs and SQLs.
We pulled some budget from branded search and spread it across non-brand campaigns, and blended efficiency improved.
But the last-touch brand report dropped, search impression share fell to 63%, and the budget went right back in with a target of 90% impression share.
Nick sees this as a change management problem. Many marketers use a multi-touch model internally for granular spend decisions while the wider organization reports on something simpler.
He admitted that moving an executive team off a model it already agreed on is hard. I don't think running two models would have saved my brand budget, since the board would still have been looking at last touch.
Column fodder and the 81%
Brand came up next, and Nick pointed to Nike as a current example of what happens when a company defunds brand.
He expects some B2B CEOs to dismiss that as a consumer problem. He thinks it applies to them too.
His argument goes back to timing. Once a buyer requests a demo, they've usually picked a favorite.
Sometimes your company is only there as column fodder: procurement tells the buyer to find two more vendors, so it has competing bids to negotiate against.
"You might get the demo request just because you're doing the performance marketing... but the brand is what's gonna help you win it," Nick said.
That led to the question I'd been waiting to ask: should marketing be commissioned like sales?
Dreamdata's data shows that 81% of the journey happens before marketing hands the account to sales. On a 100-day journey, marketing has the account for 81 days and sales for 19.
Nick expects that share to shift further toward marketing as buyers get comfortable using AI to make larger software purchases on their own.
Sales used to own what to buy, why to buy it, and how. Now, he says, it mostly owns the how.
Still, he didn't land on commission for marketing. Marketing activity sits a long way from closed revenue, and he asked whether you'd commission a product team for building a great product.
He also pointed to differences in the people who choose sales careers and marketing careers.
He does expect the conversation to happen soon. I'm unsure where it lands.
The case for tying marketing pay to revenue gets stronger as the 81% grows, and the 272-day timing problem doesn't get any easier.
Final thoughts
In Nick's experience, most of the sourcing debate traces back to how people get paid.
When AEs earn more for self-sourced deals, they fight SDRs. When a board comps on last touch, branded search gets funded to protect a report.
Dreamdata's answer was to put the whole executive team on the same ARR and retention goals.
If you're a marketing leader trying to move your company off marketing-sourced vs. sales-sourced reporting, start where Nick did.
Pull conversion rates for deals with heavy marketing involvement and for deals with little or none. Bring that comparison to your next budget conversation, alongside two or three attribution models.
As buyers do more of their research alone and with AI tools, the sales leg of the relay keeps getting shorter. Most comp plans were built for a longer one, and I expect the companies that redesign them first to stop spending so much energy on the credit game.
