Data Scientist

Data Scientist - Marketing Measurement & Causal Inference
Contract / Consultancy | Remote | Immediate start
InfluencyIQ is building a campaign intelligence platform designed to help brands understand what their marketing is actually moving across channels.
Rather than simply reporting platform attribution, InfluencyIQ connects marketing activity with search, web, commerce, CRM, social and other business signals to understand how demand is generated, progresses through the customer journey and is ultimately captured.
We are looking for an experienced Data Scientist to review, validate and formalise the statistical methodology behind our Intelligence Engine.
This is initially a focused consultancy project rather than a broad data science role.
The project
We have already developed the product and analytical framework for the Intelligence Engine, including:
campaign baselines and expected performance
event-response analysis
lagged relationships
contributor-to-outcome analysis
cross-signal pathways
overlapping marketing activity
confounder detection and treatment
repeated relationships across campaigns and entities
contributor comparison
evidence classification
multiple-testing / false-positive control
progressive intelligence where customer history is limited
structured analytical outputs for downstream AI interpretation
Your role will be to take this product framework and determine the most statistically defensible way to implement it.
You will not be starting from a blank page. We have a detailed Measurement & Analysis Framework which defines what the product needs to achieve, the analytical principles it must follow, and where specialist statistical decisions are still required.
What we need you to do
You will review the framework and:
validate or challenge the proposed analytical methodology
define appropriate statistical methods for different data and relationship types
formalise baseline methodologies and minimum data requirements
define approaches to trend, seasonality and autocorrelation
establish event-response and lag-analysis methodology
determine when Pearson, Spearman, regression, count models, time-series methods or other approaches are appropriate
recommend appropriate quasi-experimental methods where natural controls exist
define treatment of overlapping contributors and multicollinearity
establish methodology for repeated activations and entity-level learning
formalise multiple-testing and false-discovery controls
define effect-size, uncertainty and statistical-reliability requirements
establish criteria for evidence classifications such as Strong, Moderate and Limited
define when the system should return No Detectable Relationship, Insufficient Evidence or Unable to Isolate
formalise how pathway-supported contribution should be evaluated
define rules for progressive analytical maturity as additional evidence accumulates
help design a synthetic validation and backtesting framework
document the resulting methodology clearly enough for our engineering team to implement in production
You may also be asked to prototype or provide reference implementations of selected analytical methods.
A key measurement problem we are solving
A core part of InfluencyIQ is distinguishing attribution from contribution.
For example, a customer may:
see a product through a creator campaign;
subsequently search for the product;
click a paid Google result;
purchase.
Google Ads may receive direct conversion attribution because it captured the final click. InfluencyIQ needs to analyse the wider evidence to determine whether creator activity appears to have generated the upstream demand subsequently captured by paid search, without making unsupported causal or monetary attribution claims.
We are looking for someone who is comfortable solving this type of problem using incomplete, multi-source observational data.
Essential experience
You should have strong practical experience in several of the following:
statistical modelling
causal inference
time-series analysis
experimental or quasi-experimental design
event studies
regression modelling
treatment of confounding and selection bias
multiple-hypothesis testing
uncertainty estimation
repeated-measures or hierarchical modelling
observational data analysis
model validation and backtesting
You should also be comfortable translating statistical methodology into clear decision rules that software engineers can implement.
Strong Python experience and familiarity with mainstream statistical/data science libraries is expected.
Particularly relevant backgrounds
Experience in marketing analytics is useful but not essential.
We would be particularly interested in candidates with backgrounds in:
marketing measurement
econometrics
causal inference
experimentation
product analytics
marketplace analytics
advertising measurement
media effectiveness
marketing mix modelling
attribution
commercial analytics
Candidates from other sectors are equally welcome where the underlying methodological experience is strong.
Nice to have
Experience with any of the following would be useful:
Marketing Mix Modelling
Multi-Touch Attribution
incrementality testing
geo experiments
difference-in-differences
synthetic controls
Bayesian methods
adstock and saturation modelling
digital advertising platforms
GA4 / Google Ads / Meta Ads
Shopify or ecommerce data
Google Search Console
CRM data
creator / influencer marketing measurement
What we are not looking for
This is not primarily:
a dashboard / BI role
a data engineering role
an LLM engineering role
a generic machine-learning role
a paid-media optimisation role
We specifically need someone with depth in measurement, statistical inference and observational-data methodology.
Initial deliverables
The first phase is expected to produce:
A reviewed and annotated version of our Measurement & Analysis Framework.
A formal statistical methodology specification.
Recommended models, thresholds, diagnostics and decision rules.
Clear identification of any areas where the proposed methodology should change.
A validation and synthetic-testing plan.
A handover suitable for our engineering team to productionise.
There may be an opportunity for an ongoing advisory relationship as InfluencyIQ develops more advanced modelling, including incrementality, longitudinal organisation intelligence and future cross-organisation recommendation and budget-allocation models.
About InfluencyIQ
InfluencyIQ is a campaign intelligence platform designed to connect marketing activity with business outcomes across channels.
Our aim is to move beyond siloed platform reporting and simplistic last-click attribution, giving marketing teams a clearer view of what appears to generate demand, what captures it, how signals move across the customer journey and what they should do next.

Work arrangement
Yes