How Good Is Your Advice, Really?

It's possible to improve advice with better experience through feedback.

Posted Jan 05, 2018

Be it on business, health, or relationships, how often do you give advice to others? Are you a good advisor? How do you know?

Over the course of our lives, we (the authors) have benefited immensely from the advice of others. We have talked to those who had faced similar problems in life, consulted experts about their specific knowledge, and learned from their collective wisdom before making important decisions. We were (and still are) thankful for our advisors’ help and typically expressed our gratitude to them on the spot.

Yet we now realize that we actually wronged many of them.

The art and science of receiving and giving advice has rightfully garnered a lot of attention, especially in business.  Advice can be extraordinarily beneficial when provided in an appropriate manner. Garvin and Margolis’ recent HBR article provides one of the complete guidelines. And yet, even they fail to mention a key element that we noticed when we switched roles and became advisors ourselves.

A missing ingredient

Consider any successful advisory process. The advisees solicit advice, and the advisors give it. The advisees then consider the advice, act, and experience the subsequent outcomes. All good until here… Then most communication breaks down, however. They almost never provide the advisor's feedback on their specific guidance. Which parts of the advice were useful? Which were useless, or even possibly harmful? Curiously, this feedback is not a default behavior.

As an advisor, how often do you hear back from the people you advised? As an advisee, how often do you inform your advisors about the effectiveness of their various suggestions?

Advisors need such experience to refine their intuitions, beliefs, and subsequently, their future suggestions. How can they improve their advice without learning what happened? Worse, in the absence of that feedback, bad advice can calcify and persist indefinitely.

Mechanisms to improve advice

One domain where improving and refining advice are vital is medicine (literally). Put yourself in the shoes of an emergency doctor. You treat patients and send them away, almost never to receive any feedback. Did that one feel better? Was a particular approach effective? Maybe it was harmful? No clue… When another patient shows up with similar ailments the default would be to act similarly, absent any evidence to the contrary.

Emergency doctors (and physicians in general) would benefit tremendously from timely feedback and follow-ups that provide direct evidence on their treatments of individual patients (here’s some evidence from an actual emergency doctor). They would be motivated by hearing from those who benefited from their interventions. And they would be helped to adjust their behavior in light of comments by those who were less than impressed.

Unfortunately, there’s no system in place to guarantee that people would call the emergency rooms they visited to provide feedback. Follow-ups after regular visits are also problematic (check also the comments to the linked article: many people don’t follow-up because they had a problem with the physician, which actually ensures that the problem persists). One possible mechanism to incentivize patients to give feedback would be to provide discounts on fees if patients return surveys about their treatments after a few days. We are not aware of any such scheme today, or even if it has ever been tested.

Coming back to advice in business and social life, we are jokingly considering asking our advisees to deposit some collateral with us, which we’ll return when they inform us about the validity of our advice at a mutually agreed time. We think we are good advisors. But it’s time to find out if we are really as good as we think we are and improve our guidance accordingly. Providing feedback to advisors on their advice should be common practice, conventional wisdom, default behavior… We owe this much to our past advisors and future advisees.