The Essential Importance of First-Party Data

September 21, 2018      Kevin Schulman, Founder, DonorVoice and DVCanvass

Nick just sounded the warning bell about the inaccuracy of third-party data.

I want to follow up with a more fundamental question: Even if third-party data were more accurate, why would you want it?

After all, looking at the Predictably Inaccurate study from Deloitte, when data-brokers had the data wrong and even when people found out about it, only 37% opted to edit their own data. 

The top reasons they gave for not editing their data?

  • Privacy
  • No perceived value/not worth the time and energy
  • Not interested/don’t care what data they have on me

Or, in the words of a respondent:

“I’m skeptical and cautious about what could be done with this data. Even assuming the best of intentions and integrity by people who might consume this data, I cannot imagine a scenario that would also be in my or my family’s best interest. I would actually prefer less personal information about me to exist publicly. So, obscure, inaccurate, or unreliable data is what I consider to be the next best thing.”

So these data are things that at worst end-users don’t want you to have,  and at best are apathetic about or has no relevance to them.  It doesn’t sound like something worth paying to append to your file.

Why are we doing this?

Because it’s easy.

We unintentionally make efficiency the enemy of effectiveness when we, for example, hyper-focus our segmentation project on the need to score the full database with segment membership, immediately. This becomes both the end and the means as it dictates process and technical solution; a technical solution requires using 3rd party data – the really crappy, unreliable and undifferentiated stuff – to create our segments. This translates to shitty data pushed through crappy, unsubstantiated models based on false assumptions and causations, but hey, we got a score for everybody…

To cover up this original sin, consultancies (it is almost always outside agencies selling this snake oil) do an attitudinal survey that builds on this “segmented” foundation of dirty, crappy sand. Never mind that the attitudes are gross generalizations on top of a segmentation that is heavily driven by age or perhaps people who collect stamps or like red hats; always some random hodgepodge of underlying, irrelevant to supporter needs, 3rd party variables that just so happen to create the illusion of meaningful segments. Nick made one up a few weeks ago and I swear it’s indistinguishable from the tripe peddled for five and six figures.

Mindset and culture don’t really eat strategy for breakfast.  Instead they dictate what you have for breakfast and how breakfast is prepared (tactics). We need to get into the mindset of getting first-party data from donors. It’s accurate. It’s proprietary to you. It’s specific to that donor’s needs and desires. It’s a sustainable competitive advantage.

This idea of first-party data as strategy is paradoxically foreign and familiar for charities. The latter is the (poor) attempt to track supporter behavior with your organization. Charities are pretty good at tracking giving but generally horrible at tracking other behavior. Witness those 6,431 codes in a variable/field that is supposed to indicate interest except nobody knows what the codes mean, they’ve changed every time staff has changed, and most aren’t populated with anything valuable.

It’s time, as Roger has argued, for data stewardship. Part of that is getting your constituents to create their own data – first-party data.

We’ve heard the objections:

  • People don’t reply to surveys
    • Counterpoint: more people respond to surveys than to appeals and we still send those out
    • Counterpoint: think census, not sample. If you build donor Identity and Commitment level into your acquisition process and treat it for what it is – as valuable as bank details – you have 100% coverage for all newly acquired.  What about our existing supporters?  You can get coverage there, over time but focus on your real pain point – year one, if only because it comes before year 2,3,4…
  • People don’t do what they say
    • Counterpoint: They also don’t do what they did as evidenced by retention rates and 2nd gift conversion.
    • Ask a crappy survey question and you get crappy answers
    • Ask a well-constructed one on something they can reliably answer and you get insights that you simply cannot get anywhere else – period.
  • Where do we store it?
    • Counterpoint: Storing this doesn’t require symbology or speaking in an ancient tongue.  It is letters and numbers, you know alpha and numeric.
    • Counterpoint: Stop overcomplicating everything.  Standing committees at the UN spend less time deliberating on nuclear proliferation then gets spent debating how to store information in your CRM.  It is letters and numbers
  • How would we use it?
    • Counterpoint: This, in a nutshell, is where all the organizations wind up when their end game is scoring a database immediately.  No obvious, clear, simple to follow plan to use the segmentation scheme.
    • Counterpoint: if you don’t know how to use data you are collecting and storing on the CRM than stop doing it. This is not so much a regulatory comment (though that matters with GDPR), as it is a time suck one. Time is our most precious commodity, save yourself a few meetings if the question of data storage doesn’t start with how you plan to use it.

Getting the proper first-party data is of course harder.  But, if you’re going to grow your organization and succeed you must understand the “why” of your donors’ giving.

And you need to be able to cater to it.

But no third-party source is going to tell you that – you have to get it on your own.

Kevin

2 responses to “The Essential Importance of First-Party Data”

  1. Alex says:

    “Why are we doing this? Because it’s easy.” Yep. Penny wise and pound foolish (literally and metaphorically) And it sounds sexy to those (perhaps even your boss) who may not understand the uselessness – ‘oi! we’ve meta-scored every constituent against the superdupertrooper-ai-looper-quantum-materially-mechanic -model-regression-analysis-#donorlove-product and discovered delineated high propensity cluster(fucks) that will lead to exponential incremental growth and self-improving, self- sustaining…hold on, there’s a donor call on the line? Nevermind, let it go to voicemail…’

    “Mindset and culture don’t really eat strategy for breakfast. Instead they dictate what you have for breakfast and how breakfast is prepared (tactics).” <- well put Kevin.

  2. Where does most third-party data come from? First-party data!

    There’s not a single reader out there who can say that they haven’t filled out a survey with inaccurate information. It’s human nature to describe a different (better) description of our self image – on paper at least. So filling out surveys are something that a lot of people with a version of their own persona. Dating profiles have become the poster child for this type of psychoanalysis.

    So, first-party data is misleading and third-party is garbage. What’s to be done? Taking in first-party data and pushing it through a validation and curation process produces data of the highest quality – just ask hubspot. Simply using progressive profiling as a method of validation and curation is a great way to improve your data. It provides instant information and allows you to continuously change the information you capture – you may not know all the questions to ask right now. Imagine if you could cooperatively share that information!

    There’s also some value in zeroth-party data, you know, facts. How about just knowing more about the geography/location of your donors? Do they live a rural area? Urban? What does there census information tell us about them? We heard for years about how “we wish we could just add census data to our database!”. Well, you can. And it takes about 10 minutes, and it’s free.

    Did you skip second-party data on purpose? This is clearly the most misunderstood and underutilized source but that’s mostly because flexible and secure cooperative databases don’t exist… yet 🙂