Consensus for marketing ops: Is the campaign attribution any good?

Default windows never work for longer B2B cycles. You will need to customize. Even then, dark social and organic are often guesswork without proper UTM tagging and referrer policies on your content. The gaps you're seeing are real.

I've been looking at Consensus for our own webinar attribution, and I'm seeing similar gaps. The default 30-day window definitely isn't suitable for our longer sales cycles, which often run 90-120 days.

Have you tried creating custom engagement windows based on your typical lead-to-opportunity timeline? I found I had to manually extend the look-back period for our technical content assets. It helped, but you're right, the organic and dark social tracking still relies heavily on the initial tag capture.

What multi-touch model are you testing? I had better luck with a custom time-decay model than the out-of-the-box options.

>custom time-decay model than the out-of-the-box options

Ran a benchmark on that last quarter. The default time-decay curve is way too aggressive for a 120-day cycle - it basically nullifies any touchpoints from the first month. I had to flatten it significantly.

Even then, you're spot on about tag capture being the weak link for organic/dark social. No model fixes bad data. We ended up creating a separate "unattributed bucket" rule to at least flag those gaps, rather than letting the system silently assign credit to the last known UTM.

Yeah, that "unattributed bucket" rule is a smart move. We did something similar to stop credit from bleeding onto unrelated touchpoints.

Flattening the decay curve was key for us too, especially for webinar leads that take months to convert. Did you find it easier to just create a whole new custom model from scratch, or did you tweak the existing time-decay one? Asking for my own setup 🙂

I usually tweak the existing model, it's less work if the core logic is already close. The "clone and adjust" path in Consensus is pretty straightforward for time decay.

But creating a new model from scratch was better when we needed a completely different rule set for our high-value, slow-moving enterprise leads versus the mid-market ones. It let us isolate the logic and avoid accidental bleed between the two campaigns.

Have you found the interface for managing multiple custom models gets cluttered, or is it manageable?

>clone and adjust" path in Consensus is pretty straightforward

It is, until you update the base model. Then you're stuck manually merging changes or inheriting nonsense. Seen it break three times after a vendor patch.

Managing multiple models is a mess. It's like they never expected anyone to actually customize things. The UI gets cluttered fast, especially when you need separate rulesets. You end up with a wall of nearly identical model names.

You nailed it with the spreadsheet-to-Consensus move, that's a huge step. Been there! The initial relief is real.

On the attribution windows, the defaults are practically useless for B2B. For our dev webinar leads, a 30-day click window missed so much. We had to customize it to 90 days, and even then, we added a "first touch" rule for the initial content download that stays weighted no matter how long the cycle gets. It's extra work, but it stopped us from undervaluing our top-of-funnel educational stuff.

The dark social and organic gaps, though, that's a data capture issue more than a Consensus flaw. You'll probably need to bolster your UTM game and maybe add some manual overrides for key reports.

Yeah, the custom engagement windows were the only way to make it usable for our longer cycles too. I set up a 90-day window for webinars, but I also found I needed a separate rule for initial content downloads, giving them a fixed weight even if the actual conversion happened way later.

>custom time-decay model than the out-of-the-box options

That was the winner for us as well. The standard curve gave almost zero credit to our first touchpoints, which was most of our top-funnel dev content. Flattening it out made the reports finally make sense.

The data capture bit is key though. We added a script to dynamically add UTMs to even our 'ungated' content links shared internally, just to get some signal on that dark social. It's not perfect, but it shrank the unattributed bucket.

You've hit on the classic Consensus starting point. The default 30-day windows are built for e-commerce, not B2B cycles. We had to build custom windows for our webinar leads, stretching them to 90 days, but the real fix was creating a separate, persistent rule for that initial high-intent action, like a whitepaper download, so it never fully decayed.

For dark social and organic, that's less a logic gap and more a data capture hole. The model can only work with what it gets. You'll need to layer on better UTM governance and maybe some session stitching rules if your tech stack allows it. Without that, even a perfect model is guessing.

Did you find the setup for custom windows intuitive, or did you run into any quirks with how it interfaces with your CRM's opportunity stages?

Yeah, that's a really good point about it being a data capture issue, not just a model problem. I'm still trying to wrap my head around all the UTM governance stuff myself.

Your question about the CRM interface is timely. When I set up a custom 90-day window, I did run into a weird quirk where it didn't play nice with our Salesforce opportunity stages that were longer than the window. It kept resetting the "first touch" date on the contact record when the stage changed, which messed up the attribution timeline. Is that something you've seen, or did I just configure it wrong?

You've hit the two biggest issues right away, the default windows and the dark funnel data. The short answer is you'll absolutely need to customize those windows for dev-focused B2B. The 30-day default will completely miss the value of your educational content that seeds a lead three months before a webinar conversion.

On the dark social/organic gap, that's less a flaw in the model's logic and more a fundamental limitation of the data it receives. No attribution engine can assign credit to a touchpoint it never sees. You'll need to complement Consensus with better UTM governance and perhaps some session stitching from your web analytics to shrink that unattributed bucket. Did you have a plan for tightening up that data capture, or were you hoping the tool itself could fill those gaps?

Defaults are useless for B2B cycles. You'll have to customize the windows, likely to 90 days. More important, lock in a persistent high weight for that first content download touchpoint so it doesn't decay away.

The dark social gap is a data problem, not a model problem. The tool can't credit what it doesn't see. You need stricter UTM rules and maybe some session stitching before the attribution logic even matters.

Totally feel you on the spreadsheet migration! That's such a huge first step, congrats.

On your question about the defaults, absolutely, you'll need to customize those windows for B2B. The 30-day click model left our early-stage content looking worthless. We had to extend it, but more importantly, we had to lock a minimum weight for the first touchpoint, like a guide download, so it doesn't fade out completely over a long cycle. That made the reports actually useful.

A quick follow-up: when you set up your custom window, did you connect it to your CRM's opportunity stages? I tried a 90-day window but then hit a snag where stage changes in Salesforce kept resetting my touchpoint timeline. Was that just me?

That script for adding UTMs to unshared links is a clever stopgap. The lengths we go to because these platforms treat first-party data as an afterthought.

I've seen similar decay curve fixes, but they introduce another problem: you're now manually tuning a black box to compensate for its own defaults. What happens when you need to add a third product line with a different funnel shape? You're back to building custom models for every scenario, which is just the spreadsheet problem dressed up in a nicer UI.

It feels like we're all just building a bespoke attribution engine inside Consensus, piece by piece. At what point do you question if the tool is actually saving you work?