Illustrative analysis · no measured business results
Frame an analysis around a decision
DecisionName what the work should help decide
Public comparisonCompare a defined set of offers
DashboardTrack measures that support a decision
Forecast checkTest whether records support a prediction
Each route needs relevant evidence, clear assumptions and a named reviewer.
START WITH A DECISION
What do you need to understand?
The work begins with a question, the people who will use the answer and the evidence
available. A short, agreed scope sets the geography, comparison set or data sources, period
and review steps before analysis starts.
A scope we can review
Keep each kind of analysis clear.
Compare a defined set of competitors
We agree which businesses and geography to include, then review dated public sources
such as their published service pages, product descriptions and customer-facing
journeys. The comparison can cover stated offers, product or service scope, and visible
steps from first contact to purchase or enquiry. It is a bounded sample, not an
exhaustive view of a market.
Separate observations from interpretation
Notes identify what a source directly says or shows, what may follow from that evidence,
and what remains unknown. Public pages do not reveal every price, customer experience,
sales result or internal process. We do not infer market share, demand, rankings or
performance from a small comparison.
Describe activity with a dashboard
A descriptive dashboard summarises agreed measures in supplied records, such as activity
over time or the status of work. We first define each measure, its source, reporting
period and known data-quality limits. A dashboard describes the records it receives; it
does not by itself explain causes or predict what will happen.
Assess machine learning and forecasting feasibility
A forecast needs a defined outcome to estimate, suitable historical data and a way to
check predictions against data not used to fit the model. A feasibility review can
examine data coverage, consistency, time span, important changes and a sensible
evaluation approach. If the evidence is not suitable, the result may be a recommendation
not to forecast. Feasibility is not a forecast or a promise of accuracy.
WHAT TO AGREE FIRST
Boundaries, sources and review.
For competitor work, agree the geography, named comparison set, public source types and
observation dates. For reporting or forecasting, agree the business question, outcome or
measure, date range, data owner and review contact. The agreed scope records exclusions and
unresolved questions as well.
WHAT YOU CAN PROVIDE
Context and permission to use the evidence.
Share the decision to make, relevant business context and any customer-owned records needed
for the work. Data should be supplied through an agreed, appropriate channel; do not send
credentials or sensitive records in an initial enquiry. We confirm access, permitted use,
definitions and review responsibilities before working with customer data.
DELIVERABLES
Useful working material, with its limits attached.
Depending on the agreed scope, outputs may include a dated source log, comparison table,
dashboard specification or view, data-quality notes, and a written summary of observations,
inferences, unknowns and next questions. You review factual context and definitions before
the findings are treated as decision input. The work supports a decision; it cannot
guarantee a business result.
PUBLIC SOURCES
Use sources within their stated terms.
Competitor research uses accessible public material and records where and when it was
observed. It does not involve private accounts, access controls, personal data collection or
scraping content where the source or applicable terms do not permit it. Sources can change,
be incomplete or describe an offer differently from how it is delivered.
DATA AND PREDICTION
Evidence sets the limit.
Reporting depends on accurate definitions and supplied records. Gaps, inconsistent entry,
small samples or changes in how a business operates can make a measure misleading. Forecast
feasibility depends on the outcome, data volume and quality, relevant history and a fair
evaluation design. Neither a chart nor a model proves that an action caused a result.
Start with the question you need answered.
We can discuss the evidence, scope and review needed to make the work useful.