There are certain phrases that make marketers’ eyes glaze over instantly. Marketing mix modelling. Attribution frameworks. Neural networks. I know. I can almost hear you groaning already. And yet, here I am, the person at the dinner party who insists on talking about econometrics while everyone else is trying to get to dessert.
Go on, mock me if you like. Because in 2025, something genuinely exciting is happening in the world of B2B marketing measurement – and if we’re serious about proving the value of what we do, we need to pay attention.
The Problem With Old‑School Attribution
For years, UK marketers have lived with blunt tools. Last‑click attribution gave far too much credit to that one Google Ad someone happened to click on just before converting. Traditional marketing mix modelling (MMM) has been around for decades, but it relied on huge datasets, long lead times and a level of statistical wizardry that felt like it belonged in a dusty academic journal rather than a fast‑moving campaign review.
The result? B2B marketers have often struggled to defend budgets or shift spend quickly because the models just didn’t keep up with reality.
Enter the Neural Network Era
Today, we’re seeing a new generation of MMM that blends classic econometrics with machine learning. Techniques such as NNN (Neural Network‑based Modelling) and LinkedIn’s LiDDA (LinkedIn Data‑Driven Attribution) are transforming how we see ROI.
NNN uses deep learning to process vast datasets – think years of channel spend, seasonal factors, competitor activity and macroeconomic indicators – and it does so in a fraction of the time traditional MMM would take. It spots non‑linear relationships and cross‑channel interactions that a human simply wouldn’t notice.
LiDDA, meanwhile, takes the richness of LinkedIn’s first‑party data and overlays it with machine learning to show how touchpoints like sponsored content, InMail and retargeting combine to influence B2B buying journeys. For UK marketers running complex account‑based campaigns, this level of clarity is gold dust.
A Deeper Dive: NNN and LiDDA in Action
You’re still here? Brilliant. Either you really care about attribution or you’re just here for the amusement of watching me geek out. Either way, let’s dive deeper.
NNN – Neural Network‑based Modelling
Instead of relying on simple linear assumptions, NNN feeds hundreds – sometimes thousands – of signals into a machine‑learning framework that mimics the way human neurons process information.
It doesn’t just ask, “Did spend on LinkedIn drive leads?”
It asks, “How does LinkedIn spend interact with events, seasonality, macroeconomic changes, competitor activity and sales team capacity to create leads over time?”
Use Case:
Imagine you’re a software firm selling enterprise solutions. You’re investing across LinkedIn Ads, content syndication, webinars and regional trade shows. With NNN, you can model how trade‑show spend in Q1 might influence LinkedIn engagement in Q2, and how both combine to lift deal velocity six months down the line. Instead of killing a tactic too early, you can see the latent impact and justify longer‑term plays.
LiDDA – LinkedIn Data‑Driven Attribution
LiDDA is LinkedIn’s proprietary attribution solution, designed specifically for B2B buying journeys. It uses LinkedIn’s vast first‑party data to build a probabilistic map of how different touchpoints contribute to a conversion.
It moves beyond last‑touch and linear models.
It assigns fractional credit to each interaction (sponsored posts, retargeting, InMail, thought‑leadership videos) based on patterns observed across millions of journeys.
Use Case:
You’re running an account‑based marketing campaign targeting CTOs in the financial services sector. They see a sponsored video in February, download a whitepaper in March, attend a webinar in April, and finally fill out a contact form in May. LiDDA shows you that the video drove 20% of the conversion probability, the whitepaper 40%, the webinar 30% and the retargeted ads the final 10%. Suddenly, you know not just what worked, but in what combination.
Where It Gets Really Useful
Both NNN and LiDDA help answer the questions we’ve been asking for years but never had the tools to solve properly.
✅ Which channels really drive pipeline growth – and which are just noise?
✅ How do long buying cycles and multiple touches actually work together?
✅ Where can I shift spend mid‑quarter to maximise ROI without waiting for year‑end reports?
In practical terms, UK marketers are already applying these models to:
- Optimise regional spend: Retail‑tech firms using NNN to see how London vs Manchester events affect online lead gen.
- Refine content strategy: SaaS marketers feeding LiDDA insights back to content teams to prioritise whitepapers over webinars, or vice versa.
- Support budget negotiations: Marketing leaders showing finance teams hard evidence of pipeline uplift linked to multi‑channel investment.
- ABM refinement: Sales and marketing teams using LiDDA insights to tailor outreach sequences for specific high‑value accounts.
Why It Matters
Yes, it’s technical. Yes, I can hear the chuckles as I bang on about neural nets. But these tools are finally giving us a way to match B2B marketing’s complexity with equally sophisticated measurement. And for UK marketers, operating in a cautious, ROI‑driven environment, that’s not just interesting – it’s essential.
