"Garbage in, garbage out," one colleague used to say periodically, when cautioning and advising a client on making decisions. Basically he means that the integrity - validity and reliability - of information you draw on must be sound, or else any output from that information (i.e., decisions) is immediately suspect.
That said, consider the following from an article that Don Peppers posted on LinkedIn - How to Know When Data Can Be Trusted:
A recent Forrester survey of some 58,000 consumers shows that while 70% of us trust the opinions and product reviews of our friends, just 10% of us trust advertising messages. This doesn’t mean we think the ads are lying, but that they are biased. By definition, advertising is not objective. How could it be? The only reason a company invests in an advertising campaign in the first place is to recoup that investment and more in the form of product sales. Ads are designed to persuade. They are biased toward getting you to buy.
On the other hand, when you ask a friend’s opinion of a product, your friend is unlikely to have such a bias. His opinion will therefore be much more objective.
This is one reason why, as information continues to inundate all of us more and more, we will begin to rely on “social filtering” to make sense of things. What do our friends think of this, or the friends of our friends?
Social filtering allows you to enhance the objectivity of the information you encounter largely because it helps you rely on your friends’ opinions, which are less likely to be biased.
It is one thing for me to have had a friend for a long while, and I know how he or she thinks, and another thing for me to have a friend on Facebook, whom I may know only circumstantially, if at all.
Also, Facebook as a social platform is an advertising platform as well. More specifically, it is the curator of social media and is the liaison for advertisers. In this respect, I simply do not trust what they say my friends like. Again, I don't necessarily know who those friends are - in fact, I wonder how many of us actually know our friends on Facebook - and clicking Like is such an easy, convenient thing to do, that I wonder how much of us actually even read, viewed, or otherwise considered a particular post.
I may be wrong about this, but I believe that Facebook has exploited that ease and convenience to sell its social platform to advertisers. So that social filtering that Peppers hangs his trust-hat on is in itself suspect.
So ask that very question he asks, How do we know when we can trust data?
For one, trust is a very complicated thing, and it defies simplistic reasoning, such as Peppers'. Trust in itself is a human phenomenon, that is, subject to the best and the worse of what makes all of us human. To parcel it as an issue of objective versus subjective is naive, because in its essence it is some balance of both: We simply cannot know everything, about anything, at any given time.
So we must gather evidence and observations, feedback and facts, and then use our knowledge and experience to render a judgment about something. This, in brief, is an essential process of trust.
But even then it's still an imperfect process. How many times has a spouse trusted his or her partner, only to have that trust ruptured with some betrayal or another? A married couple may have been together for years, perhaps carefully nurturing a trusting relationship, but one incident can undo it in a heartbeat.
So when it comes to answering that question, I argue first that we must understand as best as we can the nature and process of trust.
Second, the issue of data itself is a very distinct one. What does it speak to, and how was it gathered? How does it relate to other data that's been gathered, and what other data needs to be gathered to address particular business matters? How much diversity is there among the types of data gathered, and how well can they be synthesized or integrated for analysis?
The question that Peppers asks is a crucial one, indeed. The correct answers are described and argued than actually obtained and executed.
Thank you for reading, and let me know what you think!
Ron Villejo, PhD