Data-driven marketing can help a business make a better marketing decision. It cannot guarantee a competitive advantage, higher return on investment or customer loyalty. Begin with one question you can act on: Which audience should receive an offer? Is a channel worth continuing? Where do people abandon an enquiry process? Define the decision before adding tools or collecting more personal information.
Updated September 2026. This is a marketing decision guide, not privacy or legal advice. Check current requirements with a qualified adviser before using personal information.
Choose a limited measure
Write down the current baseline and one measure that reflects the decision. For an enquiry campaign, that might be qualified enquiries per week and staff time required to follow up, not just clicks. Note the period, source and missing records. A small test can be useful even when it is not conclusive, provided you do not claim more certainty than the comparison permits.
Check data purpose and quality first
Ask what information was collected, for what purpose, who can access it and whether the proposed use is appropriate. B.C.’s private-sector privacy guidance explains that organizations must follow the Personal Information Protection Act (PIPA) when collecting, using or disclosing personal information. Different activities or locations may trigger other rules. For commercial emails or texts, the CRTC’s CASL guidance explains consent, identification and unsubscribe requirements, subject to applicable exceptions. A generic opt-out notice is not a substitute for checking the specific communication and its lawful basis.
Check for duplicates, stale contact information, inconsistent definitions and missing records before analysing a campaign. Do not repeat an unsourced percentage of supposedly defective records. Do not combine customer information across platforms merely because a tool makes it easy: first test the purpose, authority, necessity, access control and security of the combination. Seek qualified privacy or legal advice for your own situation.
Run one realistic data-driven marketing test
Specify the audience, message, channel, cost, staff capacity and review date. Use the least personal information needed. Compare the result with the baseline and record other plausible explanations, such as seasonality, a pricing change or different follow-up. A CRM or analytics platform may help maintain records, but the tool is not evidence that the campaign will work. Verify current privacy settings, retention options and costs before use.
Decide, then learn
At the review date, ask what changed, whether the measure is reliable and whether the cost is justified. Continue, revise, stop or investigate further. If the data cannot distinguish alternatives, say so. For example, a business might test a clearer enquiry-page message for four weeks and compare qualified enquiries and follow-up time with the preceding period. This is a hypothetical test design, not a documented client result or a promise of conversions.
For a broader framework on defining options, testing assumptions and recording a choice, see our practical business decision-making guide. Marketing data is most useful when it supports a clear decision rather than a dashboard for its own sake.
Frequently asked questions
What data should a small business collect?
Start with what is necessary for a defined decision and lawful purpose. An aggregate count may be enough when individual profiles are not needed.
Does personalization always improve results?
No. It can add privacy, implementation and trust risks. Compare results in your own setting and respect the limits of the evidence.
Can coaching confirm that a marketing data plan is compliant?
No. Coaching can help frame the decision, experiment and accountability; a qualified privacy or legal adviser should assess compliance.
Make a measured marketing decision
If you want help turning a marketing question into a plan, apply for the two-week coaching assessment. Applications are reviewed manually; acceptance and results are not guaranteed.
