Evidence-based Marketing: How your peers protect against bad marketing data


There are so many difficult decisions to make in marketing

– Which headline will perform the best?
– Which value proposition resonates most with my potential customers?
– Which call-to-action will be most effective?

This is why evidence-based marketing resonates so strongly with some marketers. As opposed to taking a random guess to answer one of these questions, why not make the decision based on prevailing evidence?

And yet, this raises another challenge. To make good evidence-based decisions, you need accurate evidence.

To help you make business decisions on a solid footing, in today’s Web clinic at 4 p.m. EDT (educational funding provided by HubSpot) – Bad Data: The 3 validity threats that make your tests look conclusive (when they are deeply flawed) – we’ll show you a few of the ways we ensure our marketing tests, and the data they produce, are valid.

But first, we asked your peers for their top marketing data quality tips. They covered a wide spectrum of marketing data approaches and use cases…

Proper planning

Bad marketing data is a result of poor project planning and management. Here are some ideas to consider when planning your next market intelligence project:

  1. Define the objective
  2. Determine if this is an in-house or outsourced project
  3. Identify the resources/set the budget
  4. Determine sample size.
  5. Establish the controllables
  6. Set a timeline

I also like to do a soft roll-out (or stagger/phase my sampling) whenever possible. This way I can adjust and make the necessary tweaks. It’s the end result that’s important.

In the end, the best protection against bad data is understanding your market. You won’t know good data from bad data if you don’t know what you’re looking at.

– Kevin Sakamoto, territory manager, Advanced Traffic Products

Validate data

First, gain consensus on what “good” data is – referring to typical data quality metrics (e.g. % completeness and accuracy). Then reach agreement on what “good enough” data is for the purpose that you intend to use it for. As an example, completely inaccurate and incomplete Date of Birth is unimportant if you’re looking at geographic distribution without an age element to demographics but is critical if you’re using it to do an age segmented mailshot.

Second, identify ways to validate data and establish data quality metrics and monitoring. Couple of easy ways to do this – data profiling of the data set to spot anomalies and doing multiple source data comparison (e.g. obtaining data from two suppliers and seeing how well they match).

Finally find ways of fixing the data if it’s wrong or not fit for the purpose.

– Gary Nuttall, Business Intelligence Analysis Team Leader, Chaucer Syndicates Ltd

Statistically sound data to reduce risk

A business must look at marketing data as a means of reducing risk.

If you are making a decision and you feel there is no risk, then you don’t need any data. If on the other hand, you are introducing a new product, the data accuracy from customized studies must be statistically sound.

At a previous company, my peers and I used the term BAD data to put this idea in perspective, with BAD being an acronym for “Best Available Data.” In essence, some data is better than no data.

If you are worried the data you have is not from a reputable source, and your gut says the opportunity needs to be validated, then you need to spend money on qualitative and/or quantitative studies that provide a reduced margin of error.

– Jeff Haltrecht, business advisor

Related Resources:

Bad Data: The 3 validity threats that make your tests look conclusive (when they are deeply flawed) – Today at 4 p.m. EDT (educational funding provided by HubSpot)

You might also like

Leave A Reply

Your email address will not be published.