Embark on the full life cycle of advanced attribution projects with our range of services. From attribution audits and incrementality tests to marketing mix modeling and in-house advanced attribution reporting, we break down the science of attribution into bite-sized pieces.
Attribution is a combination of triangulation experiments, marketing mix modeling, post purchase surveys and
last-click attribution.
The real mystery is how they come together.
Talk to us to find out.
M-Squared conducts a preliminary attribution analysis of your current media portfolio to estimate how your paid media contributes to business KPIs. A measurement roadmap and strategy, tailored for your media mix and brand objectives, is designed to reveal more detailed insights and inform future investment decisions.
[For illustrative purposes, your project timeline will be calibrated based on your project parameters]
- Prelim analysis of existing tracking and measurement systems
- Sensitivity analysis
- Gaps
- Advanced Measurement Strategy
- Measurement Roadmap
1. Prelim analysis of existing tracking and measurement systems
- Briefing of systems, access to systems
- Measurement: In-house attribution, MMM, MTA
- Tracking: Web analytics, platform reporting, surveys
- Metrics: Lower funnel events: systems, upperfunnel events: AddToCart, visits
- Dimensions: Tactics, Geo
- Sample reports
- Identifying taxonomy gaps, special cases
- Broad review of 12-18 month marketing strategy and priorities [Eg: growth targets, efficiency targets, board/exec concerns and priorities]
- Establishing scope of measurement systems in place [channel coverage, metrics coverage, dimensions coverage, measurement rigor]
- Course marketing strategy and priorities roadmap
2. Sensitivity analysis
- Channel-wise analysis of performance reported by different measurement paradigms.
- Mock investment decisions by channel.
- Sensitivity Analysis Report: Identify channels that are most sensitive to measurement paradigms
- Tracking Strenth Report: Tracking strength by data source for each channel
3. Gaps
- CTaxonomy: Aligning channel taxonomy with customer funnel and investment decisioning
- Mock investment decisions by channel
- “Current state of measurement” briefing deck for stakeholders
- Measurement solution architecture recommendations
- Taxonomy recommendations
- Tracking recommendations
- Data warehousing & BI recommendations
4. Advanced measurement strategy
- Base systems: Choice of tracking system for each channel [Web analytics, platform reporting, surveys, iSpot]
- Measurement system: Choice of channels vs Role of MMM, MTA, testing, TV attribution, iSpot
- Channel-wise measurement audit
- For a period of time (month/quarter), for each channel,
- Calculate multipliers using base tracking data and advanced measurement data
- Analyze performance and make mock decisions
- Triangulate decisions with secondary measurement source
- Identify channels that have measurement gaps and weaknesses
- Score weaknesses against marketing strategy and priorities
- Measurement solution architecture recommendations
- Channel-wise measurement stack [base, primary, secondary systems]
- Sample cross-channel dashboard
5. Measurement roadmap
- Rank order channel decisions prioritized for importance and time constraints
- Identify channels that require tracking enhancements to capture behavior.
- Identify channels that require testing and/or advanced measurement
- Testing and/or MMM roadmap
- Tracking roadmap
- Data warehousing, BI roadmap
Our agile approach to marketing mix modeling delivers results in weeks, not months, to understand impact of cross-channel media investments, non-media factors, and econometric variables on business KPIs and marketing ROI. New open source libraries combined with advancements in the underlying marketing sciences now make MMM a useful tool for quick, directional reads on customer acquisition efforts or more elaborate reports that examine the impact of full-funnel marketing efforts.
[For illustrative purposes, your project timeline will be calibrated based on your project parameters]
- Data collection
- Data Quality Analysis
- Data Quality Analysis
- Data Prep
- Modeling
- Model Interpretation & Analysis
- Finalize Model
- Triangulation Analysis
- Opportunity Analysis
- Finalize recommendations
Typical Phases in a MMM Project
- Understand marketer’s business questions and related decisions to shape modelable objectives.
- We collect data from the brand studies, consumer behavior actions from 1st party assets like site and apps, brand searches trends and other brand and consideration proxies to run qualitative and quantitative analyses and synthesize a High Value Actions (HVAs) index that could be used as a proxy for brand and consideration signals.
- We collect and shape data to support a tranche of modeling and run quality analysis process to ensure data is fit for modeling.
- We then prepare the data to build models.
- We build separate models by different objectives (Eg: By Country, By Sales Channel, by app platform: iOS app, Android app) and use iterative modeling to identify significant input features (2-3 tranches each model).
- Run a decomposition analysis (1-2 rounds for each model), repeating each model to see how to group features and get significant MMM reads (3-4 iterations).
- Finalize MMM reads and arrive at a recommended structure for brand’s MMM.
- We interpret the results and read out channel contributions and efficiencies to outline MMM enhancements roadmap.
- An analytics framework is built to assess LTV of customers sourced from marketing channels and connect it to causal channel ROI measurement
- Triangulation Analysis Overlay is used by overlaying the MMM reads and Channel LTV estimates on a triangulation analysis template and estimate causal channel ROAS based on LTV estimates.
- We identify channels that are driving high value customers profitably.
- Opportunity Analysis - where we assess channel investment opportunities that could maximize growing high LTV customers profitably.
- An fully setup infrastructure to warehouse the selected model and generate decomposition results on a weekly basis to drive in-house advanced attribution reporting.
Geo-matched market tests use first party data to test channels like Facebook, TikTok, Google Shopping, YouTube, OTT and others for incrementality and scale potential. A dedicated team will work with you to design, flight, read, and interpret incrementality tests to understand media contribution to business KPIs and the true ROI of media investments.
[For illustrative purposes, your project timeline will be calibrated based on your project parameters]
- Shape objectives
- Testing requirements, Sample datasets, Final dataset
- Market selection, First draft of selected markets, Feasibility analysis, Holdout vs Scale approach, Budgets, Decision Matrix
- Discussion, Final markets selection for flight, Signoff
- Media planning & buying Cool down
- Media flight
- Process data for reads
- Counterfactuals, Initial read, analysis, discussion
- Final read, Media mix decisions, presentation to stakeholders
Typical Phases in a Geo Match Market Incrementality Test
- Shape testable objectives.
- Design the test suited to your needs by indentifying market selection analysis and test markets
- We design test cells and run feasibility analysis.
- We then prepare final calibrated test design and run it alongside test budgets.
- Prepare decision matrix.
- Support trafficking media campaigns to test designs.
- Periodically monitor market performance and campaign execution.
- Comprehensive lift reads through lift analysis via counterfactual predictions.
- Interpreted test results, calculate multipliers, media and consumer insights, investment decisions vs decision matrix.
- New learnings agenda.
Lift tests are available to advertisers within most ad platforms to test for incrementality and scale on those individual channels. Working with your platform account team, they are typically simple to set up and run, but they tend to be opaque and tricky to interpret since they are based on proprietary data owned by the platforms themselves. Our team will help set up your tests to be interpreted in a transparent manner, compare the results with tests from other platforms, and benchmark results from other brands to validate the reads and interpret them for investment decisions.
[For illustrative purposes, your project timeline will be calibrated based on your project parameters]
- Shaping testable objectives
- Testing requirements, sample datasets, simulating for statistical significance, feasibility analysis, holdout vs scale approach, budgets, decision matrix
- Discussion, final design for flight, Signoff
- Communications with platform account team and supporting test setup
- Test flight
- Process data for reads
- Initial read, multiplier analysis, triangulation analysis, sanity analysis
- Draft interpretations and recommendations. Discussion with the project stakeholder.
- Final read, media mix decisions, presentation to broader stakeholders
Typical Phases of a Platform Lift Test
- Shape testable objectives.
- Design the test suited to your needs by indentifying campaigns to include in test.
- Design test cells for stat sig.
- Run feasibility analysis
- We then prepare final calibrated test design and run it alongside test budgets.
- Prepare decision matrix.
- Support communicating with platform account teams to trafficking the test.
- Periodically monitor test performance and campaign execution.
- Interpret test results coming out of platform reports and square off test reads with last click reporting.
- Calculate multipliers, sanity analysis, media and consumer insights, investment decisions vs decision matrix
- New learnings agenda.
See below some of the frequently asked questions about the services we offer!
If you can’t find one of your questions here, feel free to get in touch with us!
Whatever questions you may have or topics you want to cover, we would love to hear from you!
Fill out the form below or make a booking with us.
