I outline the process of improving your ideas at work. This is a non-negotiable skill for knowledge workers in any creative market. We all have to build an air-tight business case for things we believe will make a difference. If we cannot come up with better ways to satisfy our customers, build more useful products, and solve problems faster and cheaper, we will no longer be relevant.
Yet, the question of how to improve our ideas is rarely discussed.
Traditionally, ideas improvement has been filed as an information access problem. Get me access to better and more information and I’ll give you better ideas. Access is not that big a problem anymore. In the post-Internet and LLM era, information is near universally accessible today. Intelligence is not.
In fact, as I’m going to argue, more information by itself breeds a crippling disease that hasn’t been diagnosed clearly enough.That disease is confirmation bias, by the way.
More on all of this below.
A few weeks back I ran a survey and some kind folks among you suggested that I put out shorter pieces that could be read through in 5-7 minutes. So I’m trying out a crisper format. **Let me know how this one did in the comments. I’ll iterate on it. **
Let’s jump in.
The age-old way of making a recommendation at work:
Step 1: Form a view.
Step 2: Build a thesis or argument for that view.
Step 3: Gather data/information to support that argument.
Mostly, Step 3 is not that hard today thanks to the Internet. Step 1 is a breeze too because first impressions are cheap. Step 2 is about coming up with a reason, which is not that hard if you’ve ever put in the effort to make an excuse. What’s the problem, then?
There are two bugs in this algorithm. They relate to:
🐛Bug #1: How we form beliefs
Belief-making is backward. Research shows that we form beliefs and then construct reasons that back them up. We don’t first find reasons and then stitch them together into a belief.
This means that sometimes we lack useful information to construct an argument for our decisions. Some other times, we do but we let it pass because we’re busy looking for reasons that justify our first impressions. Sounds a lot like steps 2 and 3 in our algo, doesn’t it?
Note: I promised to try a compact format so I’m not quoting Daniel Kahneman on the belief-making process, but for those interested, here’s a transcript.
🐛**Bug #2: How more information makes you more confident but not accurate **
The time spent looking for reasons that justify our first impressions is time taken away from evaluating the set of information that exists and is available. When we let our gut guide us before we have gathered all relevant and available information, we canter toward forming a generalized overall impression based on a narrow dimension judged early.
There’s a story I share with my decision-making community when I tell them about confirmation bias. I’ll skip the story but here are the facts.
Researcher Paul Slovic ran an experiment on horse-racing handicappers (professionals who analyze horse-racing data to pick potential winners for races) where he continued to feed them more and more data points with each round. The handicappers picked whichever data points they wanted and made predictions.
As you can see below, the predictions didn’t get better. Which was bad enough but what made it worse was that their confidence shot up with more data points. More confidence meant bigger bets meant bigger losses because the bet accuracy had flatlined.
>>>>> gd2md-html alert: inline image link here (to
images/image25.png). Store image on your image server and adjust path/filename/extension if necessary.
(Back to top)(Next alert)
>>>>>
Remedy?
We can’t do much about how we form beliefs. If Danny Kahneman and fellow behavioral scientists, after years of studying the human mind, admit to the inescapable influence of our brain machinery, then what reason do we have to think we are beyond those limitations?
And what can we do about confirmation bias? Reject information? Tell our colleagues and teams not to give us data because that’ll have a perverse effect on our thinking?
Here’s how to improve our ideas and beliefs.
1️⃣Farm for dissent
2️⃣Socialize idea
3️⃣Get buy-in
Note: The names for steps 1 and 2 are taken from a stage in The Netflix Innovation Cycle. For more literature and to find out what the Netflix folks are drinking, read Reed Hastings’ book No Rules Rules.
If you look at the steps, you’ll notice that the sequencing allows for your confidence to build. That makes sense. You wouldn’t want to follow a process where you start out wildly sure and end up with knots in your stomach. It is in how you’re letting your confidence grow that makes the difference.
💡You’re not working harder alone to gather more information and letting your belief deepen in the process. You’re shifting your focus from _more information _to more perspectives.
You’re doing that, and I’m going to get into the how of it next, by leaning on others and tapping into their wisdom.
Farm for dissent
At the beginning, when you perhaps should be the least confident about your idea, share it with others in a document with a permissive brief: tell me what you make of it.
Ask those you’ve reached out to to vet your idea. They could just leave their comments in the document, along with everyone else. Allowing people to rate it on a scale (like Netflix does, between -10 and +10), along with giving comments, makes for a better device for capturing the intensity of feeling without relying on the writing skills of the feedback-giver.
Though the step is called farm for dissent, you’re in a way searching for endorsement as well. Should you find encouraging inputs for your idea, it is early validation that you’re onto something.
Should you add the numbers? Should you calculate the median? The raters are not calibrated among themselves, so I wouldn’t trust the median score. The scale is a useful model until overused. The point is less about measurement, more about evaluation.
Whom should you reach out to? Don’t reach out en masse. You’re not looking for a people’s choice vote. Those in the orbit of the decision or idea you’re considering are generally a good qualifier. I’ve also found going to colleagues cross-functionally and those one level lower to be helpful in showing me my blind spots.
Socialize your idea
Farming for dissent can give you a good spread of things to tinker with, unless the message is a clear yes or no. Once you’ve smoothed out the edges, it is time to then go deeper—stress-test your idea by introducing it in some depth to others. This is socializing your idea.
Set up 1-on-1 chats with a smaller group of people with the objective of curating their opinions on your idea. There are at least two changes in this step from the previous one: the medium of communication switches from written to oral AND the conversation deepens. Both changes go hand in hand.
This step produces a rather useful by-product. It improves your communication. In the process of outlining your proposal to colleagues, you work out how to make the pitch, how to ask good questions, what signs to look for in your audience.
Socializing your idea takes time because you’re speaking to a bunch of people and assimilating their input. It may also cost you money if you’re actually testing anything out with users.
Get buy-in
Let’s say you’ve farmed for dissent and socialized your idea. The hard yards are done. Still, the final stretch is often slippery. When you’ve to push your idea to your boss and your boss’s boss, you need more than data points. You need to get buy in.
This step is something consultants do well. Take your idea up to your key stakeholders and ask for their opinion. It uncovers roadblocks you may not have seen, like shifted priorities that may scupper your plan, AND it turns potential adversaries into collaborators. Now you have someone higher up who is trusted in the system evangelizing your idea.
Here’s a 2-min video I shot some time back breaking down the buy-in step.
This piece started with how to improve your ideas and ended with how to save them from dying. Implicit in this message is the fact that ideas die all the time. It is not enough to find a way to improve your idea if you cannot get people to bet on it.
I’ve had peers and bosses who are good at one or two of the steps of the idea-improvement algorithm. Or if they apply it through and through, they do so only for a while. If you’re thinking of going down this path, you’re already in the top quartile of your peers.
Years ago, I had convinced people high up in my organization of the idea of building a marketplace where scientists and scientific journals could find each other. I had done so without taking enough care to stress-test my idea. I was able to push the idea out but the marketplace didn’t take off. Since, I’ve considered the experience a lesson in the value of following a process to fine-tune my ideas.
Right now I’m in the early stages of scoping out another idea and I would be thankful if you took a few minutes to fill out this short survey. It’s for everyone but hand on heart I think it’ll resonate most with those in mid-career.