Growth Dynamics

Is your Go-to-Market strategy falling short of delivering the returns you need? 
Are misaligned teams causing delays and missed revenue opportunities? 
Is marketing seen as a cost center instead of a growth engine?

And what about your attribution models? Are they too opaque to trust, or are you relying on outdated multi-touch attribution for demand generation, leaving you with incomplete insights that lead to poor investment decisions?

I Solve These Growth Barriers:

- Misaligned Teams Slowing You Down? I bring marketing, sales, and finance into alignment, ensuring all teams work together towards shared goals that drive revenue impact.

- Opaque Attribution Hurting Your Investment Decisions? I implement clear, data-driven attribution models that give you full transparency on what’s driving real results across all your marketing channels.

- Struggling to Scale Predictably? I optimize your Go-to-Market strategy to unlock new growth opportunities and ensure your resources are allocated where they generate the highest ROI.

How I Help You Succeed:

1. Align your revenue teams to drive maximum performance.

2. Set up transparent attribution tools that show exactly which channels drive demand, whether you’re using multi-touch or something else.

3.Scale strategically with custom performance insights that ensure your investments fuel measurable growth.

Ready to Turn Your Go-to-Market Strategy into a Revenue Powerhouse?

Let’s talk. I specialize in solving opaque attribution problems, aligning teams, and creating demand generation strategies that produce predictable, scalable growth.

Founder

Oracle

Principal Data Strategist

•	Committed to hands-on, inspiring leadership - promoting growth, teamwork, and innovation within teams.
•	Collaboratively established the roadmap with data producers and users to align objectives with senior leadership vision.
•	Fostered robust relationships with Oracle's senior stakeholders for enhanced collaboration and addressing needs.
•	Initiated and managed the execution of high-impact internal data initiatives and products, ensuring compliance with Oracle Security and Compliance Standards, and navigated the approval process with key stakeholders including executives, legal, compliance, and IT teams.
Team key projects:
-	Implemented the new Media Data Model following the Company strategy reorganisation: implementing new data onboarding using APIs, metric definitions, and robust data governance across Search, Social, Digital, Online comparison, Content Syndication and Emails.
-	Developed an automated data quality and control tool, significantly reducing error rates from 10% to 0.1% for both qualitative and quantitative metrics.
-	Leveraged automated behaviour analysis to uncover key insights into user journeys, and optimisation opportunities using Multi-Touch Attribution.
-	Developed a model to estimate the impact of marketing initiatives on organic traffic
-	Devised a Mixed Media Modelling optimisation tool, employing non-linear regression analysis to maximise ROI
-	Pioneered a Pipeline/ARR forecasting tool utilising Boosting Trees (Cat Boost) for accurate long-term predictions, integrating GTM initiatives and external factors impacting demand with a 20% accuracy on year-long funnels.
-	Developed measurement frameworks to define the contribution of each channel on a single strategy
-	Designed a Demand Generation and Capture model to drive revenue growth and business success.
-    Lead and contact scoring model
-      API connections to Google Ads, Campaign Manager, Google Search Console, Linkedin

Manager of Media Analytics

• Successfully built and led a diverse, remote team focused on Data Science, Data Engineering, and Analytics projects, driving revenue growth. 
• The objective of my team is to support Senior Management, Strategy and Channels in the optimisation of marketing strategy towards:
 - creation of tools (Forecasting , Marginal ROI )
 - data access and reporting improvements ( data and taxonomy control tools, additional data requests)
 - Performance investigation and creation of insights on live campaigns

Manager of Data Science - Paid Media

Skai

• Investigating opportunities for insight creation and marketing optimisation using a wide variety of data
• Leading key multi-person project agenda on “digital user intent” using NLP and clustering
• Leading the knowledge agenda of the team, mentoring to close knowledge gaps
• Ensuring team alignment with market through various meetings with Sales, Client Services and Strategists
• Evaluation of data quality and behaviour to ensure smooth integration into models using Machine Learning
• Investigation of performance anomalies to provide solutions or model adaptations for clients and stakeholders
• Meeting clients to assess business needs, and to recommend ideas and testing frameworks
• Participation in client pitches and new client on-boarding.
• Creation and maintenance of Python tools and libraries. Key contributor to transfer of R to Python tools packages
• Designing dashboards to automate and speed up performance investigations
• Introducing new team process to improve synergies and actively contributed to tools standardisation

Product Data Scientist

Societe Generale Corporate and Investment Banking - SGCIB

•	Managed product life-cycle events and the correct integration and pricing of initiated deals into the systems
•	Controlled and monitored trade amendments and cancellation of pending trades; troubleshooted P&L issues
•	Supervised compressions and initial margin optimisation on Rates and FX 
•	Actively engaged with the development team to increase the automation of life cycle events
•	Improved KPI and workflow process controls with scripts created in Python:
    o	A text-mining model to cross-check bookings with client contracts.
    o	Daily automated reports to ensure workflow quality control and chase interested parties required actions

Exotic Credits Trade Support

Pordum Investments

•	Designed and implemented statistics-based strategies for day trading and swing trading
•	Undertook quantitative market analysis to provide forecasting and market trend analysis 
•	Supervised and managed large market research projects to successfully find new tradable edges
•	Undertook many pieces of statistical research to improve quantitative understanding of market behaviour 


Main research areas:
- Movement-based charts trading strategies
- Multi-timeframe trading strategies 
- Profitability of trading market sentiment (retail positioning contrarian)
- Absolute currencies’ strength trading strategies

FX Trader

BNY Mellon

•	Responsible for the management of high risk Corporate Actions (up to 800 clients) and the follow up of VIP accounts (with assets under custody up to 5 billion USD)

•	Managed cash accounts and FX deals to ensure liquidity requirements for Corporate Events processing

•	Ensured accounts reconciliation by investigating Asset, Cash and FX related issues 

•	Engaged in process optimisation and risk management improvement by developing macros,  working closely with IT for the creation of new systems, and updating procedures

•	Participated in the impact analysis of changes in market specificities and regulations and submitted changes to processing

•	Avoided losses for the Bank and clients thanks to excellent problem solving aptitudes within extreme time limits and very good negotiation skills with brokers and sub-custodians

•	Successfully managed the contingency tests with the Corporate Actions Teams around the world

•	Conducted in-depth analysis of the implementation of the TARGET systems in the European custodial market and defined business opportunities (in Product Management team)

Operations Associate - Market Expert

Corporate Actions - Summer Trainee

•	Generated a deep analysis of the implementation of the TARGET systems in the European custodial market and researched new opportunities for the Bank

Assistant Product Manager (Internship)

Fortis Insurance Belgium

•	Designed the product brochure of 2008, presented it to the Board and enabled its acceptance

•	Proposed product improvements to increase their competitiveness and instilled pressure on the different sales departments

Marketing Intern

AXA Investment Managers

•	Followed trainings on asset valuation and asset allocation, and answered to client queries

•	Analysed customers needs to increase sales productivity

Asset Management Intern

Data Strategy and Optimisation expert - More Revenue, Less Cost

𝐆𝐫𝐨𝐰𝐭𝐡 𝐢𝐬 𝐒𝐭𝐚𝐥𝐥𝐢𝐧𝐠—𝐇𝐞𝐫𝐞’𝐬 𝐖𝐡𝐲 (𝐀𝐧𝐝 𝐇𝐨𝐰 𝐭𝐨 𝐅𝐢𝐱 𝐈𝐭)

Most companies think they have a marketing problem. In reality, they have a misalignment problem—between sales, marketing, and finance.

🔴 Sales says marketing leads don’t convert.
🔴 Marketing says sales doesn’t follow up properly.
🔴 Finance is questioning why CAC keeps rising.

𝐒𝐨𝐮𝐧𝐝 𝐟𝐚𝐦𝐢𝐥𝐢𝐚𝐫?

📉 Companies over-focus on short-term lead generation and under-invest in brand; shrinking their pipeline before sales even starts.

📉 They track what’s easiest to measure (MQLs, last-click attribution) instead of what actually drives revenue (brand, demand, buyer trust).

📉 They optimise for conversion, not consideration; forgetting that only 5% of buyers are in-market at any given time, and 80% will pick a brand they already know before they even start their buying journey.

𝐖𝐡𝐚𝐭 𝐈 𝐃𝐨

I help B2B companies with $12M+ sales & marketing budgets turn misaligned teams into predictable revenue machines by fixing their growth strategy, marketing measurement, and go-to-market execution.

✔ Marketing shouldn’t just generate leads—it should shorten sales cycles, improve conversion rates, and increase retention.
✔ Attribution should guide better decisions—not justify what already happened.
✔ Sales & marketing should work together—not fight over credit.

𝐖𝐡𝐲 𝐖𝐨𝐫𝐤 𝐖𝐢𝐭𝐡 𝐌𝐞?

📊 Built and scaled GTM strategies for companies managing $100M+ in marketing budgets across Tech, Finance, Automotive, and more.

📊 Ex-Oracle Principal Data Strategist. Designed advanced measurement systems that aligned marketing investment with revenue impact.

📊 Background in capital markets & quantitative finance. I don’t just look at marketing & sales in isolation; I help companies make financially sound go-to-market decisions that improve margins, efficiency, and long-term value.

𝐈𝐟 𝐘𝐨𝐮’𝐫𝐞 𝐑𝐞𝐚𝐝𝐲 𝐓𝐨:

✔ Stop chasing cheap leads that don’t convert
✔ Build a marketing function that improves outbound success
✔ Reduce CAC and increase pipeline velocity
✔ Align sales & marketing so your sales team actually hits quota

📩 𝐋𝐞𝐭’𝐬 𝐭𝐚𝐥𝐤.

👉 https://calendly.com/charlie-7y0/30min

𝐈 𝐜𝐡𝐚𝐥𝐥𝐞𝐧𝐠𝐞 𝐚𝐧𝐲 𝐂𝐄𝐎, 𝐂𝐅𝐎, 𝐨𝐫 𝐂𝐌𝐎 𝐭𝐨 𝐩𝐫𝐨𝐯𝐞 𝐭𝐨 𝐦𝐞 𝐭𝐡𝐞𝐲 𝐡𝐚𝐯𝐞 “𝐧𝐨 𝐛𝐮𝐝𝐠𝐞𝐭” 𝐟𝐨𝐫 𝐛𝐫𝐚𝐧𝐝 𝐚𝐰𝐚𝐫𝐞𝐧𝐞𝐬𝐬.

𝐇𝐞𝐫𝐞’𝐬 𝐰𝐡𝐚𝐭 𝐈 𝐬𝐞𝐞 𝐚𝐠𝐚𝐢𝐧 𝐚𝐧𝐝 𝐚𝐠𝐚𝐢𝐧:
👉 Companies spend 80–90%+ of their marketing budget on conversion and retargeting.
👉 Their revenue curves always look the same. I’ve never seen an exception.


𝐓𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦
20–30% of that conversion spend sits in the death zone.

Cut it tomorrow, and your revenue will barely move. That money isn’t working. It’s wasted.


𝐓𝐡𝐞 𝐜𝐮𝐥𝐩𝐫𝐢𝐭: 𝐀𝐭𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧.

Attribution gives you the false impression that you know what is going on. 

But it only measures the bottom of the funnel. 

Whether you use it to optimise campaigns or evaluate your marketing team, it always leads to the same outcome: wasted spend at the bottom of the funnel, oversaturating the small group of buyers already in market and willing to buy from you.

For high-ticket purchases, 80% of buyers will choose a brand they already know before their buying journey even begins.

If you only focus on consideration and retargeting, 𝐲𝐨𝐮’𝐫𝐞 𝐟𝐢𝐠𝐡𝐭𝐢𝐧𝐠 𝐨𝐯𝐞𝐫 𝐚 𝐭𝐢𝐧𝐲 𝐟𝐫𝐚𝐜𝐭𝐢𝐨𝐧 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐓𝐨𝐭𝐚𝐥 𝐀𝐝𝐝𝐫𝐞𝐬𝐬𝐚𝐛𝐥𝐞 𝐌𝐚𝐫𝐤𝐞𝐭.


𝐓𝐡𝐞 𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧
Take that dead 20% and reinvest it in creating the buyers of tomorrow. Something retargeting and conversion initiatives cannot do.

Suddenly, you unlock the growth you’ve been chasing for years. Because brand is the only marketing asset that compounds over time.

The real risk isn’t “spending on brand.”
The real risk is starving your brand and over-feeding the death zone.

𝐒𝐭𝐨𝐩 𝐟𝐞𝐞𝐝𝐢𝐧𝐠 𝐭𝐡𝐞 𝐝𝐞𝐚𝐭𝐡 𝐳𝐨𝐧𝐞. 𝐒𝐭𝐚𝐫𝐭 𝐟𝐞𝐞𝐝𝐢𝐧𝐠 𝐭𝐡𝐞 𝐛𝐫𝐚𝐧𝐝.

💬 If you suspect part of your spend is stuck in the death zone, DM me. I’ll help you identify it!

Charlie de Thibault

Turn Wasted Budget Into Profitable Growth.

𝐖𝐡𝐲 𝐘𝐨𝐮 𝐍𝐞𝐞𝐝 𝐭𝐨 𝐁𝐚𝐜𝐤𝐭𝐞𝐬𝐭 𝐘𝐨𝐮𝐫 𝐌𝐌𝐌 𝐚𝐧𝐝 𝐖𝐡𝐲 𝐅𝐫𝐞𝐞𝐳𝐢𝐧𝐠 𝐂𝐨𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐭𝐬 𝐢𝐬 𝐌𝐚𝐥𝐩𝐫𝐚𝐜𝐭𝐢𝐜𝐞

What's the point of a coefficient of an MMM if it doesn't produce relationships you can actually rely on?

🔎 𝐖𝐡𝐚𝐭 𝐛𝐚𝐜𝐤𝐭𝐞𝐬𝐭𝐢𝐧𝐠 𝐫𝐞𝐚𝐥𝐥𝐲 𝐭𝐞𝐥𝐥𝐬 𝐲𝐨𝐮
By running your model on rolling windows, you discover whether channel contributions remain consistent or collapse as soon as the window shifts. 

When they collapse, the model is telling you something important: this channel is weak, noisy, or collinear. 

Don't trust this coefficient as stable evidence.
Backtesting exposes which channels you can rely on and which ones you can't.

✅ 𝐖𝐡𝐞𝐧 𝐬𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲 𝐢𝐬 𝐠𝐫𝐞𝐚𝐭 𝐧𝐞𝐰𝐬
If contributions stay consistent across rolling windows, the channel has robust signal

The model is learning from the data, not relying on forced priors. You can act on these coefficients with high confidence.

Stable contributions are the gold standard because they show your model is strong and doesn't need heavy-handed prior tweaks to prop it up.

🚫 𝐖𝐡𝐲 𝐟𝐫𝐞𝐞𝐳𝐢𝐧𝐠 𝐜𝐨𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐭𝐬 𝐢𝐬 𝐦𝐚𝐥𝐩𝐫𝐚𝐜𝐭𝐢𝐜𝐞
When faced with unstable channels, many practitioners take a shortcut. 

They lock ROI at a "nice" number, fix adstock half-lives, and hard-code saturation curves. It makes the outputs look tidy, but it's dishonest. 

The instability hasn't disappeared, you've just hidden it.

🎭 𝐖𝐡𝐲 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦𝐬 𝐚𝐫𝐞 𝐢𝐧𝐜𝐞𝐧𝐭𝐢𝐯𝐢𝐬𝐞𝐝 𝐭𝐨 𝐟𝐫𝐞𝐞𝐳𝐞
It's not always malice, but incentives matter. 

Client trust pressure means swings in contributions make results look unreliable, so freezing keeps clients happy.

Commercial incentives drive this too. Platforms are paid to prove their value, and unstable results undermine that narrative. 

Communication challenges compound the problem because saying "we don't know for sure" is harder than handing over a clean chart.

But the cost is real: hidden uncertainty, overconfidence in weak channels, and misleading budget allocations.

Sometimes the honest answer is simply: "This channel doesn't show a stable contribution yet. Either the signal is too weak, or it overlaps too much with other activity." That's not failure, it's valuable information.

⚖️ 𝐁𝐞𝐰𝐚𝐫𝐞 𝐨𝐟 𝐓𝐢𝐠𝐡𝐭 𝐏𝐫𝐢𝐨𝐫𝐬

Tight priors can be powerful as they bring domain knowledge and stabilise weak signals. 

But there’s a big difference between guiding the model with justified information and forcing stability to hide uncertainty. Always be transparent and justify your use of priors

🧠 𝐖𝐡𝐲 𝐢𝐧𝐬𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲 𝐡𝐚𝐩𝐩𝐞𝐧𝐬
Low spend creates low signal, leaving insufficient data to identify an effect. 

High correlation between overlapping channels blurs attribution.

Incrementality tests using correlation models ( geo, DiD) will only work with your biggest channels.

MMM and Incrementality tests are essential measurement tools, but there's a critical constraint that affects their effectiveness: the signal-to-noise ratio. 

These models work well when the signal is strong enough to rise above the noise.

This means that those models work well for:
👉 bigger channels
👉 short term effect

Imagine you want to test a small channel representing 10% of your media mix. Unless this channel drives ridiculously high ROI, the signal won't surpass the noise created by the other channels and other external influencers.

This is how you end up with an unsuccessful incrementality test for a channel which is in fact very promising.

It also drastically reduces your ability to run small tests, which kills your capacity to build a test-and-learn culture where teams can experiment and discover what actually works.

MMM is at high risk to do the same and is usually visible by the gigantic confidence intervals and the low stability when running a backtest on smaller channels. This is why the use of priors is so important to communicate what the model cannot see by itself (as long as you don't overuse it!).

This is why we approach things differently.
1. 𝐖𝐞 𝐬𝐩𝐥𝐢𝐭 𝐭𝐡𝐞 𝐟𝐮𝐧𝐧𝐞𝐥.In addition to revenue we look at earlier signals such as traffic, item viewed, items add to cart, pricing page visits
⚠️ 𝘉𝘦 𝘤𝘢𝘳𝘦𝘧𝘶𝘭 𝘰𝘯 𝘰𝘯𝘦 𝘵𝘩𝘪𝘯𝘨 𝘵𝘩𝘰𝘶𝘨𝘩! 𝘐𝘵 𝘪𝘴 𝘯𝘰𝘵 𝘣𝘦𝘤𝘢𝘶𝘴𝘦 𝘺𝘰𝘶 𝘪𝘯𝘤𝘳𝘦𝘢𝘴𝘦 𝘴𝘪𝘵𝘦 𝘷𝘪𝘴𝘪𝘵𝘴 𝘵𝘩𝘢𝘵 𝘪𝘵 𝘪𝘴 𝘢𝘭𝘸𝘢𝘺𝘴 𝘨𝘰𝘰𝘥! 𝘠𝘰𝘶 𝘯𝘦𝘦𝘥 𝘵𝘰 𝘥𝘳𝘪𝘷𝘦 𝘴𝘪𝘵𝘦 𝘷𝘪𝘴𝘪𝘵𝘴 𝘵𝘩𝘢𝘵 𝘤𝘰𝘯𝘷𝘦𝘳𝘵 𝘴𝘰 𝘺𝘰𝘶 𝘴𝘩𝘰𝘶𝘭𝘥 𝘮𝘰𝘥𝘦𝘭 𝘵𝘩𝘦 𝘳𝘦𝘭𝘢𝘵𝘪𝘰𝘯𝘴𝘩𝘪𝘱 𝘣𝘦𝘵𝘸𝘦𝘦𝘯 𝘵𝘩𝘦𝘴𝘦 𝘴𝘪𝘵𝘦 𝘷𝘪𝘴𝘪𝘵𝘴 𝘢𝘯𝘥 𝘳𝘦𝘷𝘦𝘯𝘶𝘦 𝘰𝘷𝘦𝘳 𝘵𝘪𝘮𝘦 𝘢𝘯𝘥 𝘦𝘯𝘴𝘶𝘳𝘦 𝘪𝘵 𝘥𝘰𝘦𝘴 𝘯𝘰𝘵 𝘥𝘦𝘨𝘳𝘢𝘥𝘦.

2. 𝐖𝐞 𝐬𝐭𝐢𝐥𝐥 𝐮𝐬𝐞 𝐠𝐡𝐨𝐬𝐭 / 𝐏𝐒𝐀 𝐭𝐞𝐬𝐭𝐬. Yes, this means less control on our side but there are some requests you can make to the vendors to ensure the credibility of the test.

3. 𝐖𝐞 𝐥𝐨𝐨𝐤 𝐚𝐭 𝐪𝐮𝐚𝐥𝐢𝐭𝐚𝐭𝐢𝐯𝐞 𝐝𝐚𝐭𝐚 It is a great source of information where correlations fail.
_____________________________________________________
Hey, I'm Charlie and we assist enterprises to optimise more than $150 million of marketing spend annually. We specialise in long consideration journeys and brands with offline sales. If you need help, reach out!

You can't make this stuff up. You absolutely can't rely on raw traffic data for programmatic campaigns.

Invalid traffic usually hides in sessions, but it shows up in higher-intent actions.
And it’s not even hiding well. You don’t need a complex or expensive bot detector. With a single SQL query, you can already spot a lot of it.

Just pull your GA data by device, city, and hour; it will show up.

The patterns are always the same:
- Same channels (programmatic is the worst)
- Same devices/OS (Linux, Android)
- Same countries (often the ones you don’t target)

Fraudsters trigger higher-intent traffic to manipulate algorithms and sell bot traffic.

If you’re not running this exercise, I can guarantee the “reported data” you’re using is heavily inflated.
__________________________________________________
👋 Hey, I’m Charlie. At Growth Dynamics, we help enterprises optimise more than $150M in marketing spend every year. We specialize in long consideration journeys and brands with offline sales. If you need help, reach out.

𝐌𝐌𝐌 𝐚𝐧𝐝 𝐈𝐧𝐜𝐫𝐞𝐦𝐞𝐧𝐭𝐚𝐥𝐢𝐭𝐲 𝐜𝐚𝐧𝐧𝐨𝐭 𝐦𝐞𝐚𝐬𝐮𝐫𝐞 𝐛𝐫𝐚𝐧𝐝’𝐬 𝐠𝐫𝐞𝐚𝐭𝐞𝐬𝐭 𝐟𝐞𝐚𝐭𝐮𝐫𝐞: 𝐢𝐭𝐬 𝐥𝐨𝐧𝐠-𝐭𝐞𝐫𝐦 𝐩𝐨𝐰𝐞𝐫.

At first, brand feels like it’s giving little in return. But over time, it becomes your strongest asset: lowering acquisition costs, building preference, attracting better talent, and enabling scale at higher margins.

Short-term spend, on the opposite, rents mental availability. You pay each time to be noticed.

Think of it like running: every run makes the next one easier, and you can go further. Stop for a while and yes, you’ll lose some performance but you’re still ahead of someone starting from scratch.

Brand works the same way. It fades if ignored, but with consistent delivery( in product, service, and messaging) it builds durability and makes growth easier year after year.

𝐓𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦
There’s a strong misconception that MMM can track the long-term impact of brand. It can’t.

Not only because correlation models fail to detect diluted brand signals (why MMM often over-attributes retargeting if you don’t control for it)… but because MMM is locked to past spend and past revenue. It cannot see the equity you’re building today that fuels tomorrow’s buyers.

𝐀𝐭𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧 𝐦𝐚𝐤𝐞𝐬 𝐢𝐭 𝐰𝐨𝐫𝐬𝐞
It gives you the false impression that you know what’s going on, but in reality it only tracks the bottom of the funnel.
That’s where you end up overspending, chasing short-term efficiency, and starving the brand that creates tomorrow’s buyers.

𝐌𝐲 𝐫𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐭𝐨 𝐜𝐥𝐢𝐞𝐧𝐭𝐬
👉 𝐁𝐞 𝐫𝐮𝐭𝐡𝐥𝐞𝐬𝐬 𝐢𝐧 𝐨𝐩𝐭𝐢𝐦𝐢𝐬𝐢𝐧𝐠 𝐭𝐡𝐞 𝐛𝐨𝐭𝐭𝐨𝐦 𝐨𝐟 𝐭𝐡𝐞 𝐟𝐮𝐧𝐧𝐞𝐥, because attribution, MMM and incrementality are accurate there and you can quickly find the point of diminishing marginal returns.

👉 Give brand breathing room in the short term. Because today you can only measure a fraction of its effect, and if you optimise only on what’s visible, you’ll massively underinvest in the thing that drives future growth. Don't fall into "If we can't measure , it did not happen" trap.

👉 𝐓𝐫𝐚𝐜𝐤 𝐒𝐡𝐚𝐫𝐞 𝐨𝐟 𝐒𝐞𝐚𝐫𝐜𝐡 across platforms to measure your demand relative to competitors. It’s one of the best leading indicators of brand strength and your “share” in the category.

𝐁𝐫𝐚𝐧𝐝 𝐢𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭 𝐛𝐮𝐢𝐥𝐝𝐬 𝐨𝐰𝐧𝐞𝐫𝐬𝐡𝐢𝐩 𝐨𝐟 𝐦𝐞𝐧𝐭𝐚𝐥 𝐚𝐯𝐚𝐢𝐥𝐚𝐛𝐢𝐥𝐢𝐭𝐲. Once you own it, the benefits show up in economics: margins, loyalty, talent, and scale.

𝐓𝐡𝐞 𝐃𝐮𝐚𝐥 𝐎𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞 𝐋𝐢𝐦𝐢𝐭: 𝐭𝐡𝐞 𝐡𝐢𝐝𝐝𝐞𝐧 𝐟𝐥𝐚𝐰 𝐢𝐧 𝐏𝐌𝐀𝐗, 𝐀𝐒𝐂, 𝐚𝐧𝐝 𝐞𝐯𝐞𝐧 𝐲𝐨𝐮𝐫 𝐌𝐌𝐌.

I wanted to make this a video but Black Friday prep got in the way. So here it is as a post instead.

A model, by definition, has a 𝐬𝐢𝐧𝐠𝐥𝐞 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐭𝐚𝐫𝐠𝐞𝐭.

- For MMM it's minimizing prediction error (maximizing likelihood).
- For PMAX or ASC it's maximizing platform-defined conversion proxies.

𝐘𝐨𝐮 𝐜𝐚𝐧'𝐭 "𝐛𝐚𝐤𝐞 𝐢𝐧" 𝐝𝐮𝐚𝐥 𝐨𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞𝐬 𝐢𝐧𝐭𝐨 𝐚 𝐬𝐢𝐧𝐠𝐥𝐞 𝐦𝐨𝐝𝐞𝐥. 𝐌𝐨𝐝𝐞𝐥𝐬 𝐰𝐢𝐥𝐥 𝐚𝐥𝐰𝐚𝐲𝐬 𝐜𝐨𝐥𝐥𝐚𝐩𝐬𝐞 𝐭𝐡𝐢𝐬 𝐢𝐧𝐭𝐨 𝐰𝐡𝐢𝐜𝐡𝐞𝐯𝐞𝐫 𝐨𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞 𝐠𝐞𝐭𝐬 𝐭𝐡𝐞𝐦 𝐜𝐥𝐨𝐬𝐞𝐬𝐭 𝐭𝐨 𝐭𝐡𝐞𝐢𝐫 𝐭𝐚𝐫𝐠𝐞𝐭.

𝐇𝐞𝐫𝐞'𝐬 𝐭𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦:
PMAX and ASC narrow into high-intent audiences because that maximizes their target metric: conversions. 

That looks efficient at first, but it kills reach and erodes incrementality. Why? You're just harvesting demand that was already there.

That's why when you test those channels, the incrementality is higher at the beginning than at the end. They scoop up fresh intent first, then the pool dries up.

The problem gets worse with short conversion windows. Systems trained on 1-day or 7-day windows chase quick wins and give up on higher-value, longer-cycle buyers. 

This is temporal discounting: algorithms prioritize immediate rewards over long-term value.

MMM does the same thing. It overweights short-term signals that are clear, while underweighting long-term effects that are fuzzier. 

The model chases what's easiest to explain: the short term.

𝐓𝐡𝐞 𝐫𝐞𝐚𝐥 𝐩𝐫𝐨𝐛𝐥𝐞𝐦? 𝐌𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐡𝐚𝐬 𝐭𝐰𝐨 𝐨𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞𝐬:
👉 Build awareness and demand.
👉Convert that demand into revenue.

The solution isn't forcing dual objectives into one model. It's recognizing the single-objective bias and managing around it:

For PMAX/ASC: Don't rely on them for your entire strategy. They're high-intent driven; you need other campaigns to build awareness and demand.

For MMM: Acknowledge the limit; it will always chase short-term signals over long term signals

𝐓𝐫𝐞𝐚𝐭 𝐢𝐭 𝐥𝐢𝐤𝐞 𝐚 𝐩𝐨𝐫𝐭𝐟𝐨𝐥𝐢𝐨 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐩𝐫𝐨𝐛𝐥𝐞𝐦:
You need both short-term and long-term activities. They have to be measured differently, with different models. Then you need to figure out how much to allocate to each.

That's the dual objective limit. The model can't solve it;  𝐲𝐨𝐮 𝐡𝐚𝐯𝐞 𝐭𝐨 𝐦𝐚𝐧𝐚𝐠𝐞 𝐢𝐭.

🤝 Or I can manage it for you. DMs are open🤝

I start every client onboarding with a MyTelescope analysis. I’m sharing the template here and why I use it.

myTelescope lets you track share of search and branded queries across many platforms.

𝐅𝐢𝐫𝐬𝐭, 𝐢𝐭 𝐬𝐡𝐨𝐰𝐬 𝐢𝐟 𝐰𝐞’𝐫𝐞 𝐢𝐧 𝐚 𝐠𝐫𝐨𝐰𝐢𝐧𝐠 𝐨𝐫 𝐬𝐡𝐫𝐢𝐧𝐤𝐢𝐧𝐠 𝐜𝐚𝐭𝐞𝐠𝐨𝐫𝐲. 

This external factor heavily impacts "marketing performance". If your category is growing fast, much of the performance comes from rising demand as opposed to marketing only which is something attribution will always miss. The opposite is also true.

𝐒𝐞𝐜𝐨𝐧𝐝, 𝐢𝐭 𝐫𝐞𝐯𝐞𝐚𝐥𝐬 𝐡𝐨𝐰 𝐭𝐡𝐞 𝐜𝐨𝐦𝐩𝐚𝐧𝐲 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐬 𝐰𝐢𝐭𝐡𝐢𝐧 𝐢𝐭𝐬 𝐜𝐚𝐭𝐞𝐠𝐨𝐫𝐲. 

Even if overall demand is strong, a declining share of search means you’re leaving money on the table. From the outside, the business may look good  but the reality is different.

𝐓𝐡𝐢𝐫𝐝, 𝐈 𝐥𝐨𝐨𝐤 𝐚𝐭 𝐭𝐡𝐞 𝐬𝐩𝐢𝐤𝐞𝐬. 

Companies that rely on attribution often struggle to measure the impact of upper-funnel campaigns. Showing how these “campaigns that didn’t drive anything” actually helped them dominate their category sparks powerful discussions. Clients are usually more open to testing upper-funnel campaigns when they see this data.

𝐅𝐨𝐮𝐫𝐭𝐡, 𝐈 𝐝𝐢𝐯𝐞 𝐢𝐧𝐭𝐨 𝐤𝐞𝐲𝐰𝐨𝐫𝐝𝐬. 

Which ones are increasing? Which competitors dominate? By comparing product-level share of search, we can see who’s winning specific battles in the category.

Finally, branded queries are absolute goldmines. They reveal what’s really working in the upper funnel;  the part that measurement and attribution tools often miss.

My Telescope Analysis

𝐇𝐞𝐫𝐞’𝐬 𝐚 𝐯𝐞𝐫𝐲 𝐬𝐢𝐦𝐩𝐥𝐞 𝐰𝐚𝐲 𝐭𝐨 𝐬𝐩𝐨𝐭 𝐛𝐨𝐭 𝐭𝐫𝐚𝐟𝐟𝐢𝐜 𝐢𝐧 𝐆𝐀4. 𝐘𝐨𝐮𝐫 𝐝𝐢𝐬𝐩𝐥𝐚𝐲 𝐜𝐚𝐦𝐩𝐚𝐢𝐠𝐧𝐬 𝐚𝐫𝐞 𝐟𝐮𝐥𝐥 𝐨𝐟 𝐢𝐭!

GA4’s UI is terrible,  I agree. But …

𝐓𝐡𝐞 𝐁𝐢𝐠𝐐𝐮𝐞𝐫𝐲 𝐞𝐱𝐩𝐨𝐫𝐭 𝐢𝐬 𝐚 𝐠𝐨𝐥𝐝𝐦𝐢𝐧𝐞.
If you haven’t set it up yet, that’s my first recommendation. The possibilities with GA4 data are endless, especially if you have a longer consideration journey where correlation-based measurement models fall short.

𝐒𝐭𝐞𝐩 1: 𝐆𝐞𝐭 𝐭𝐡𝐞 𝐝𝐚𝐭𝐚
Pull the following fields into BigQuery:
DATE
HOUR
SOURCE_MEDIUM
CAMPAIGN
COUNTRY / METRO
DEVICE_CATEGORY
BROWSER / OPERATING_SYSTEM
MOBILE_DEVICE_BRAND / MOBILE_DEVICE_NAME
PAGE_VIEWS
ADD_TO_CART


𝐒𝐭𝐞𝐩 2: 𝐒𝐞𝐭 𝐭𝐡𝐞 𝐭𝐡𝐫𝐞𝐬𝐡𝐨𝐥𝐝 𝐟𝐥𝐚𝐠
Look at the add-to-cart ratio:
A ratio > 1 means more add-to-carts than page views:  a massive red flag.

For a more data driven approach, check your brand search campaigns ratio 

These typically show the highest ratio, and even then it’s closer to 0.5, not 1. If you don't have a brand search campaign, Direct is a good proxy too.


𝐒𝐭𝐞𝐩 3: 𝐀𝐧𝐚𝐥𝐲𝐳𝐞
Most problematic campaigns will be display.
I’ve seen campaigns where 75% of traffic was flagged.

The usual culprits?
Devices running Linux or Android builds.

Suspicious mobile brands for a US audience like Xiaomi, Infinix, Tecno, Realme RMX and … Samsung. Less suspicious but I keep seeing that guy being flagged

𝐒𝐭𝐞𝐩 4: 𝐆𝐨 𝐟𝐮𝐫𝐭𝐡𝐞𝐫 𝐰𝐢𝐭𝐡 𝐅𝐨𝐮𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬
This method helps you to better grasp the problem size, but it only catches the most obvious bots.

For real analysis, you need a dedicated tool like FouAnalytics.


𝐖𝐡𝐲 𝐭𝐡𝐢𝐬 𝐰𝐨𝐫𝐤𝐬
Fraudsters know your campaigns optimize for conversions.
So they trigger add-to-cart events to fool the algorithm into thinking they’re sending valuable traffic. You waste the budget -> they get paid.

𝐖𝐡𝐲 𝐀𝐧𝐝𝐫𝐨𝐢𝐝
It is much easier to create a bot on an Android device as opposed to an Iphone. 



𝐁𝐨𝐧𝐮𝐬 𝐩𝐨𝐢𝐧𝐭𝐬:
This also reveals what devices drive intent and purchases.

For example, if most of your high-intent traffic and conversions are on iOS and Mac, should you keep running that campaign that drives 80% of Android traffic

In high school, my math teacher said I was a moron and I needed brainwashing. 

10 years later, I graduated with an MSc in financial mathematics from one of the most prestigious French Business School.

Now I do math for a living.

There is still much more that schools should do for neurodivergent people.

Reading this visual of some research shared by Elsa Torres, EdD Student, MSc Childhood Youth, FRSA, NPQH broke my heart and validated so many of my own school memories at the same time.

These aren’t just statistics - they’re the daily reality for millions of neurodivergent students who are told they’re “not trying hard enough” when the system itself is causing them distress.
What strikes me most is how different the pain points are, yet how similarly devastating the impact can be. Those of you who know me well, know I’m both AuADHD and left school at 15. Looking at the below, I get some clarity as to why 😬

For autistic students, it’s often about the environment and social confusion. The fire alarm that wasn’t announced. The group project where no one explains the unspoken social rules. The sensory overwhelm that gets labeled as “behavioral problems.”

For ADHD students, it’s frequently about relationships and misunderstanding. The teacher who takes fidgeting as disrespect. The “boring” tasks that feel impossible to focus on. The constant message that if you just “tried harder,” everything would be fine.

The part that gets me:
Both groups are often incredibly capable, creative, and intelligent. But when the delivery method doesn’t match how their brains work, they’re made to feel like failures instead of recognizing that the system needs adjusting.

To educators reading this: Small changes can have massive impact. Visual schedules, movement breaks, clear expectations, and understanding that different doesn’t mean deficient.

To parents: Your advocacy matters. Trust your instincts when something isn’t working for your child.

To my fellow neurodivergent adults: If school was hard for you, it wasn’t because you weren’t smart enough. The research is clear - the system wasn’t designed for how your brilliant brain works.

What would have made the biggest difference in your school experience?

#neurodiversity #education #autism #adhd #inclusion #advocacy #research

No alternative text description for this image

Use a VPN to enjoy LinkedIn without ads. They recently changed something...


I always used a VPN but still got targeted in my country. 

Probably because LinkedIn was using the country listed on my profile.

But in the past few days, that’s no longer the case. Now I get ads based on my VPN country.

Here’s the trick: pick a random country far away that no one targets (since advertising on LinkedIn is expensive), and enjoy the peace.

Branding Builds Assets
Ads decay; hacks fade. Brand equity grows in the customer’s mind, even while you sleep. That’s why lasting companies obsess over it, it keeps paying dividends decades later.
Follow for no BS growth lessons

𝐓𝐡𝐢𝐬 𝐢𝐬 𝐡𝐨𝐰 𝐰𝐞 𝐩𝐚𝐫𝐭𝐢𝐚𝐥𝐥𝐲 𝐬𝐨𝐥𝐯𝐞 𝐭𝐡𝐞 𝐥𝐨𝐧𝐠 𝐜𝐨𝐧𝐬𝐢𝐝𝐞𝐫𝐚𝐭𝐢𝐨𝐧/𝐨𝐟𝐟𝐥𝐢𝐧𝐞 𝐦𝐞𝐚𝐬𝐮𝐫𝐞𝐦𝐞𝐧𝐭 𝐜𝐡𝐚𝐥𝐥𝐞𝐧𝐠𝐞 𝐟𝐨𝐫 𝐛𝐨𝐭𝐡 𝐬𝐦𝐚𝐥𝐥 𝐚𝐧𝐝 𝐥𝐚𝐫𝐠𝐞 𝐜𝐚𝐦𝐩𝐚𝐢𝐠𝐧𝐬 𝐚𝐧𝐝 𝐠𝐞𝐭 𝐝𝐢𝐫𝐞𝐜𝐭 𝐟𝐞𝐞𝐝𝐛𝐚𝐜𝐤

Measuring long consideration journeys is particularly difficult for revenue attribution because the signal gets completely drowned in the noise. 

Plus, it's nearly impossible to do this for smaller channels because they get lost in the noise of bigger channels, which drastically reduces the capacity to run small-scale tests.

The funnel in the image shows a very classic setup I see regularly. Social and shopping drive a lot of your top-of-funnel activity but completely disappear at the bottom of the funnel.

On the other side, most revenue comes from SEO, direct traffic, brand search, and offline purchases; channels that drive very little to none upper-funnel activity.

𝘚𝘪𝘥𝘦 𝘯𝘰𝘵𝘦 𝘰𝘯 𝘚𝘌𝘖: A lot of that traffic is also branded SEO traffic. It's a classic error to attribute all of that performance to SEO, and this is why you need to use Google Search Console to properly separate branded from non-branded organic traffic.

What simply happens is that 𝐛𝐮𝐲𝐞𝐫𝐬 𝐝𝐨𝐧'𝐭 𝐜𝐨𝐧𝐯𝐞𝐫𝐭 𝐭𝐡𝐫𝐨𝐮𝐠𝐡 𝐭𝐡𝐞 𝐜𝐡𝐚𝐧𝐧𝐞𝐥 𝐰𝐡𝐞𝐫𝐞 𝐭𝐡𝐞𝐲 𝐝𝐢𝐬𝐜𝐨𝐯𝐞𝐫𝐞𝐝 𝐲𝐨𝐮. 

And MTA won't solve this because it's extremely difficult to stitch touchpoints together, leading to most top-of-funnel activity being disregarded.

What we can do is estimate revenue by activity and build a mid-funnel mixed media model to re-attribute revenue to top-of-funnel activity.

The great thing is that this also works at the campaign level and provides direct feedback. 

𝐋𝐚𝐮𝐧𝐜𝐡 𝐚 𝐜𝐚𝐦𝐩𝐚𝐢𝐠𝐧 𝐭𝐨𝐝𝐚𝐲, 𝐚𝐧𝐝 𝐲𝐨𝐮'𝐥𝐥 𝐬𝐭𝐚𝐫𝐭 𝐠𝐞𝐭𝐭𝐢𝐧𝐠 𝐫𝐞𝐚𝐝𝐬 𝐭𝐨𝐦𝐨𝐫𝐫𝐨𝐰 solving the big MMM granularity and incrementality fast feeback problem.

𝐍𝐨𝐰 𝐭𝐡𝐞𝐫𝐞 𝐚𝐫𝐞 2 𝐜𝐚𝐯𝐞𝐚𝐭𝐬:
👉 𝘋𝘢𝘵𝘢 𝘤𝘭𝘦𝘢𝘯𝘪𝘯𝘨 𝘪𝘴 𝘤𝘳𝘪𝘵𝘪𝘤𝘢𝘭. 
You need to extensively clean your website traffic data, which is full of bots. Some are good (like crawlers), some are bad. You cannot use website traffic directly from GA4.
👉 𝘞𝘦𝘣𝘴𝘪𝘵𝘦-𝘥𝘳𝘪𝘷𝘦𝘯 𝘤𝘩𝘢𝘯𝘯𝘦𝘭𝘴 𝘰𝘯𝘭𝘺. 
This only works for channels driving activity to your website. Apps like TikTok tend to drive much more indirect traffic. For those, we run mid-funnel MMMs and track mid-funnel activity lift when running incrementality tests.

We're pivoting to focus exclusively on B2C businesses with long consideration journeys and offline sales components. 

Many measurement tools break down when customer journeys exceed 4 weeks or include offline sales. 

We solve this through a holistic approach that leverages multiple data sources, each optimised for different funnel stages.

We're excited about this focus because B2C companies with these complex journeys tend to be highly motivated to solve measurement challenges.

They understand the impact of getting measurement right rather than settling for pretty dashboards. 

They're more sceptical of oversimplified attribution and more willing to embrace the nuance their business actually requires.

Working with driven, solution-oriented teams like these creates momentum for meaningful change.

_________________________________________________
Hey, I'm Charlie. We help enterprises optimise over $150 million in marketing spend annually, specialising in long consideration journeys and brands with offline sales. If you need help, reach out!

𝐓𝐡𝐢𝐬 𝐢𝐬 𝐰𝐡𝐲 (𝐚𝐧𝐝 𝐡𝐨𝐰) 𝐈 𝐮𝐬𝐞 𝐰𝐞𝐛𝐬𝐢𝐭𝐞 𝐭𝐫𝐚𝐟𝐟𝐢𝐜 𝐭𝐨 𝐩𝐥𝐚𝐧 𝐁𝐥𝐚𝐜𝐤 𝐅𝐫𝐢𝐝𝐚𝐲 𝐛𝐮𝐝𝐠𝐞𝐭 𝐚𝐥𝐥𝐨𝐜𝐚𝐭𝐢𝐨𝐧.

And this doesn’t only work for Black Friday / Cyber Monday.  It applies to any event you run during the year.

When you look at your website data, you start to see different levels of intent.

𝐀𝐛𝐨𝐮𝐭 𝐬𝐢𝐱 𝐭𝐨 𝐟𝐨𝐮𝐫 𝐰𝐞𝐞𝐤𝐬 𝐛𝐞𝐟𝐨𝐫𝐞, 𝐢𝐭𝐞𝐦 𝐯𝐢𝐞𝐰𝐬 𝐜𝐥𝐢𝐦𝐛. People are just discovering. They’re browsing, seeing ads, comparing options.
. They wonder if they’ll get a good deal, and your early campaigns are fueling that curiosity.

𝐓𝐡𝐞𝐧, 𝐚 𝐛𝐢𝐭 𝐥𝐚𝐭𝐞𝐫, 𝐚𝐝𝐝 𝐭𝐨 𝐜𝐚𝐫𝐭 𝐬𝐭𝐚𝐫𝐭𝐬 𝐭𝐨 𝐫𝐢𝐬𝐞. By this stage, people have discovered what they want, they’re saving their wishlists, and intent is much stronger.

𝐀𝐧𝐝 𝐟𝐢𝐧𝐚𝐥𝐥𝐲, 𝐨𝐧 𝐭𝐡𝐞 𝐞𝐯𝐞𝐧𝐭 𝐝𝐚𝐲 𝐢𝐭𝐬𝐞𝐥𝐟, 𝐩𝐮𝐫𝐜𝐡𝐚𝐬𝐞𝐬 𝐬𝐩𝐢𝐤𝐞.
All that pent-up demand converts in one short window.

𝐇𝐞𝐫𝐞’𝐬 𝐭𝐡𝐞 𝐜𝐡𝐚𝐥𝐥𝐞𝐧𝐠𝐞:
You can’t use revenue as your guide for budget allocation because the revenue only appears at the end. 

A lot of teams are aware of this and do spend before the event, but they rarely have a rigorous approach.

👉 𝐓𝐡𝐚𝐭’𝐬 𝐰𝐡𝐞𝐫𝐞 𝐰𝐞𝐛𝐬𝐢𝐭𝐞 𝐭𝐫𝐚𝐟𝐟𝐢𝐜 𝐜𝐨𝐦𝐞𝐬 𝐢𝐧.
By looking at year-on-year site actions, you can see exactly when item views begin to climb and when add to cart starts moving. Those signals bring rigour into your budget allocation process.

𝐓𝐡𝐞 𝐫𝐞𝐬𝐮𝐥𝐭?
A structured way to allocate spend in the 4–6 weeks leading into Black Friday not just the weekend itself.

You are welcome! Follow for more! 🚀

Why Most Brands Fail
They chase tactics without a point of view. If customers can’t say who you are in a sentence, spend is wasted. Define the promise, repeat it relentlessly, and let everything else go.
Follow for clarity that converts.

The Bouquet Test 
Every association is a “flower” in your brand bouquet, partners, products, experiences. One rotten flower can taint the whole bunch. Overwhelm negatives with consistent positives and prune fast. Follow for brand hygiene.

𝐌𝐨𝐬𝐭 𝐦𝐚𝐫𝐤𝐞𝐭𝐞𝐫𝐬 𝐬𝐭𝐫𝐮𝐠𝐠𝐥𝐞 𝐰𝐢𝐭𝐡 𝐭𝐡𝐢𝐬: 𝐇𝐨𝐰 𝐝𝐨 𝐲𝐨𝐮 𝐚𝐥𝐥𝐨𝐜𝐚𝐭𝐞 𝐬𝐩𝐞𝐧𝐝 𝐰𝐡𝐞𝐧 𝐨𝐧𝐞 𝐜𝐚𝐦𝐩𝐚𝐢𝐠𝐧 𝐝𝐫𝐢𝐯𝐞𝐬 𝐭𝐫𝐚𝐟𝐟𝐢𝐜 𝐭𝐨 𝐦𝐮𝐥𝐭𝐢𝐩𝐥𝐞 𝐩𝐫𝐨𝐝𝐮𝐜𝐭𝐬?

Ever struggled with spend allocation when running MMM across multiple product lines?

𝐇𝐞𝐫𝐞'𝐬 𝐭𝐡𝐞 𝐜𝐡𝐚𝐥𝐥𝐞𝐧𝐠𝐞: 
You're running an ecommerce brand with multiple product categories -  each with unique AOVs, customer journey lengths, seasonality, and margins. 

🚩 Throwing everything into one bucket for measurement? 🚩 

That's a recipe for inaccurate models across ALL products. When you average two different products, your model will be wrong for both.

Aggregation into product groups is beneficial as long as they behave similarly.

𝐌𝐲 𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧: 𝐆𝐀4 𝐓𝐫𝐚𝐟𝐟𝐢𝐜-𝐁𝐚𝐬𝐞𝐝 𝐀𝐥𝐥𝐨𝐜𝐚𝐭𝐢𝐨𝐧

I use GA4 traffic data to allocate spend. If I see that a campaign drives 80% of its traffic to Product A, I allocate 80% of that spend to Product A.

Why page views, not conversions? Many visitors return later and purchase through direct traffic, branded SEO, or branded paid search. 

Page views capture the initial intent-driving touchpoint.

𝑰𝒎𝒑𝒐𝒓𝒕𝒂𝒏𝒕 𝒄𝒂𝒗𝒆𝒂𝒕:This only works with campaigns actually driving website traffic.

𝐎𝐭𝐡𝐞𝐫 𝐀𝐭𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧 𝐎𝐩𝐭𝐢𝐨𝐧𝐬
Fortunately, many platforms now offer matching capabilities:
- CTV and direct mail can match order IDs when you send them your conversion feed
- Connected TV platforms increasingly offer household-level matching

𝐎𝐭𝐡𝐞𝐫 𝐕𝐚𝐥𝐢𝐝𝐚𝐭𝐢𝐨𝐧 𝐌𝐞𝐭𝐡𝐨𝐝𝐬
Incrementality testing: You'll often see different lift rates by category, validating the need for separate allocation
Creative analysis: Ad copy promoting specific products provides directional guidance for allocation

𝐄𝐱𝐚𝐦𝐩𝐥𝐞 𝐢𝐧 𝐀𝐜𝐭𝐢𝐨𝐧
Campaign "Glow Up Your Routine" → 65% traffic goes to skincare subscription pages, 35% to makeup products → 65% spend allocated to skincare subscription, 35% to makeup products in MMM model.

This approach has dramatically improved our model accuracy and given us actionable insights for each product line.

Most Influencer Impact Is Indirect
 Stop judging creators by clicks. Most buyers see content, don’t click, then Google you later and convert elsewhere. If you only track clicks, you’re ignoring most of the revenue. Measure view-through and branded search lifts to see the real picture.
 Follow for evidence-based measurement.

The Funnel Analysis Edge
You don’t always need a fancy tool. Website funnel data already shows where traffic comes from and where it drops off. It’s fast, reactive, and great for quick learning.

Branding = Credibility
Most buyers aren’t in-market today and most already know who they’ll choose when they are. Brand earns you the shortlist before the RFP, demo, or search. Build credibility now or fight for scraps later.
Follow for more.

Taplio

Charlie de Thibault's Linkedin Analytics

's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Charlie de Thibault's Linkedin Analytics

's Best Posts .css-qqfgvy{font-size:var(--chakra-fontSizes-md);}(last 30 days)

Want to drive more opportunities from LinkedIn?

Charlie de Thibault's Linkedin Analytics

's Best Posts (last 30 days)