The DOBlog

Tag: market research

  • Finding Red Herrings or Missing a Trick?

    This post is written by Colin Boylan, an independent market research professional based in Wicklow, Ireland with extensive experience in Market Research in pharma and other industries in the UK and Ireland. In this post, Colin explains how the quality of the population sample used in a market research study can have significant effects on the quality of the findings. His post was inspired by recent posts here and here about “Golden Databases“. I’m glad to give Colin a chance to try his blogging chops out and I hope visitors here enjoy reading his insights in to information quality and market research.

    Finding Red Herrings or Missing a Trick?

    For most businesses there are major advantages to investing money in doing direct research with your customer base   In theory it’s a ready built list of people who are familiar with your business – so they can speak with authority on their experience as your customer.
    The value of customer research to business should be by now fairly obvious, but there’s an old saying in research (and elsewhere) – “garbage in, garbage out”. The insights built off the data
    generated from your customer list is only as relevant as the list of people you ask to participate in the research.
    However if, for example, they are lapsed customers then researching them is going to give you a picture of what your past customers wanted from you (unless these people are the focus of your research of course).   Is this the same as what your present customers want?  And if you are looking for why past customers stopped dealing with you and use a list full of current  customers you end up with either few people able to answer the questions you set or …worse….data from people who shouldn’t have answered the question – which leads to another scenario.
    Picture an important piece of research done with a list of past and present customers mixed in together with no way to tell who is who.  Do current and ex-customers differ in their wants and needs from your business?     I don’t know – and neither do you.   So how useful are any insights generated from this research?  Not being able to separate these two groups gives rise to two potential scenarios.  Either the excess numbers in there are throwing up ‘clear’ results that are not applicable to your current customers or the combination of both bodies is adding noise which stops you uncovering real insights about the customers you’re interested in – you’re either finding red herrings or you’re missing a trick!
    I’ve used just one scenario here to make a point that can be applied to lots of customer data stored by companies – be it incorrect regional information, incorrect gender, you can add whatever block of data is relevant to your own company here and the story is the same.   If the data is not accurate then any use it is put to suffers.

    For most businesses there are major advantages to investing money in doing direct research with your customer base In theory it’s a ready built list of people who are familiar with your business – so they can speak with authority on their experience as your customer.

    The value of customer research to business should be by now fairly obvious, but there’s an old saying in research (and elsewhere) – “garbage in, garbage out”. The insights built off the data generated from your customer list is only as relevant as the list of people you ask to participate in the research.

    However if, for example, they are lapsed customers then researching them is going to give you a picture of what your past customers wanted from you (unless these people are the focus of your research of course). Is this the same as what your present customers want? And if you are looking for why past customers stopped dealing with you and use a list full of current customers you end up with either few people able to answer the questions you set or …worse….data from people who shouldn’t have answered the question – which leads to another scenario.

    Picture an important piece of research done with a list of past and present customers mixed in together with no way to tell who is who. Do current and ex-customers differ in their wants and needs from your business? I don’t know – and neither do you. So how useful are any insights generated from this research? Not being able to separate these two groups gives rise to two potential scenarios. Either the excess numbers in there are throwing up ‘clear’ results that are not applicable to your current customers or the combination of both bodies is adding noise which stops you uncovering real insights about the customers you’re interested in – you’re either finding red herrings or you’re missing a trick!

    I’ve used just one scenario here to make a point that can be applied to lots of customer data stored by companies – be it incorrect regional information, incorrect gender, you can add whatever block of data is relevant to your own company here and the story is the same. If the data is not accurate then any use it is put to suffers.

  • Golden Databases – another quick return

    I just received an email from an information quality tool vendor. It was sent to an email address I had provided to them in my capacity as a Director of the IAIDQ as part of registering for events they had run.

    The opening line of the email reads:

    I’m writing to you as a follow-up your recent telephone conversation with an [name of company deleted] representative.

    Two small problems

    1. I haven’t had a telephone conversation with any representative from this company regarding any survey or anything else recently. (I did meet one of their Dublin based team for lunch about 6 weeks ago – does that count?)
    2. The personal data I provided to them was not provided for the purpose of being emailed about surveys. (But at least they have an opt out).

    I’m going to take a look at the survey but I bet you the €250 raffle prize for participants that my responses will be statistically irrelvant.

    For a start, the survey is about the importance of the investment in data in IT planning. I’ve never worked in the IT organisation of any business. I have been on the business side interacting with IT as a provider of services to me.

    Also, as a Director of the IAIDQ and as someone trying to set up an SME business, I am basically hands on in all aspects of the business and implemenation of systems (I was working at 2am this morning doing a facelift on my company website and working on a web project for the IAIDQ). So, my responses will be misleading from a statistical point of view.

    It looks like the company in question:

    • Had an email address for me.
    • Knew that my former employer was a customer (customer #1 for this particular company’s DQ offering)
    • Forgot that I’d told their Dublin team that I’d left
    • Had failed to update their information about me.
    • Have recorded a contact with me but have recorded it incorrectly.

    Should I respond to the survey?  Will my responses be meaningful or just pointless crap that reduces the quality of the study?

    I’m goint to ask a friend of mine to write a guest post here on survey design for market research and the importance of Information Quality.