Return on Nice

My Uber rating is 4.96. Here's what I spent a year building because of it.

Hi all. This edition of The New New is breaking from my typical format. Next month, it will return with roundups and more.

For this issue, we’re going deep into a big obsession of mine — an obsession I decided to build something around.

So, fair warning: nerdfest ahead.

I would love for you to join in and let me know what you think of this one.

Let’s wonk out!

---

The Conduct Quotient

Businesses know what we spend and what we cost them. Almost none of them know what it’s like to work with us. Here’s a proposal I built so good people and good businesses can find each other.

Have you ever checked your Uber rating?

Mine is 4.96. And yes, it bugs me that it isn’t a 5.0, because I think I’ve been pretty good to those drivers.

Uber built that number so a driver could know something about a passenger before pulling up. It’s one of the only scores in my life that measures whether I’m decent to deal with.

And scores like that feel useful right about now.

Pew found that 47 percent of Americans think people have gotten ruder in public since the pandemic. Only 9 percent think we’ve gotten more polite. (I feel it every time someone plays a video out loud on their phone in public, like the rest of us signed up to watch too.)

So here’s what I keep coming back to:
Wouldn’t it be something if businesses could recognize the people who treat their staff well?

And then, wouldn’t it be even better if I could carry that “he is good to work with” recognition to a different, new-to-me business?

Those questions have been bugging me for a long time. Over the past year, I’ve been exploring this more and more.

Below is what I explored and what I built as my proposed solution. Something new for all businesses, especially service-centric businesses, could adopt. Something that could benefit all of us who believe life is better when good people find good people.

What businesses act on today

Today’s reward and recognition systems are built on the data that businesses have handy:

• What I spend

• How often do I come back

• Which tier have I reached

• What I returned, and what I disputed

• Whatever a data broker sold them about my household

Loyalty programs turn spending and frequency into points and status. Internal systems flag fraud, chargebacks, and property damage.

All of it is either financial or demographic.

None of it is what I’m actually like to work with.

There are a couple of exceptions.

Uber lets drivers set a minimum rider rating. Airbnb hosts review guests after a stay. That’s roughly it. And those ratings stay locked where they were earned. My Uber history can’t help me at a hotel. My Airbnb reviews mean nothing to an airline.

What “good to work with” actually means

Those ratings have another challenge, and this one took me longer to spot: they only look at half of what matters.

Being good to work with covers two different things.

1. The human half is how I treat the people serving me. The gate agent, the front desk, the server, whoever picks up when I call support. This is like what Uber does now with their rating.
   

2. The operational half is whether I held up my end. Did I show up? Did I pay? Did I take care of what I borrowed? Did I do what I said I would?

Together, those two help paint a clearer picture of how someone interacts with a business.

Someone can be furious about a canceled flight and still be completely fair to the gate agent trying to fix it. Someone else can be perfectly pleasant at the counter and bring the rental car back with a dent in it.

Businesses should be able to recognize the people who handle both well.

And those people should benefit from that recognition.

And that brings me to… The Idea

I am calling the idea Conduct. And within it is the proposal for a new rating/scoring system I’m calling the Conduct Quotient (CQ).

A CQ is built on the two halves I just described, each with its own measurement system (and its own wonky name and acronym, so hang in there a bit as we get nerdy here).

The human half:
Net Conduct Score (NCS)

The human half runs on a single question that a business will ask an eligible employee:

“Based only on how this customer interacted with you and others, would you want to serve them again?”

If that sounds a lot like a Net Promoter Score, bingo, because NPS is where this whole idea started. NPS asks customers how they feel about a company. This asks a company’s people how they feel about a customer. Same idea, but reversed and tweaked.

The question is also deliberately narrow. It doesn’t ask whether I was cheerful, or agreeable, or easy. It asks how I treated people. Someone can be direct, frustrated, even in open disagreement, and still score well. (In a robust website that defines the specifications behind this measurement, I went deep into the list of things the score will not measure.)

The data record half:
Verified Conduct Index

Now, let’s flip to the second part of CQ, the record side that is built from what a business already knows about each customer: Did the booking happen? Was the bill paid? Did the property come back in one piece?

It carries a rule I care about a lot: it only counts responsibilities I actually accepted and could control. So it leaves out:

• My spending

• How often have I called support

• The company’s own bad day (like a canceled flight)

A chargeback resolved in my favor doesn’t count against me. A rough rating collected during a system-wide meltdown gets treated as what it was. (This measurement also has a detailed specification backing it up.)

Two Halves → One Score

Put the two halves (the two scores above) together, and we get one number, on a 1.00 to 5.00 scale.

That is what makes up the Conduct Quotient (CQ).

Then we can look at these scores in different ways:

• for one specific relationship with a single company

• for a whole industry (like travel or restaurants/dining out)

• for a person overall (computed across every company interaction)

There are additional rules and specifications for how these scores are calculated.

For example, both halves have to qualify before any score appears, so being lovely to the staff can’t cover for never paying, and a spotless payment record can’t cover for how you treated someone. Plus, older interactions count for less than recent ones, and a pattern counts for more than one bad afternoon. And when there isn’t enough to go on, it says “building history” instead of inventing a number.

Now imagine sharing your score

So, if we each now have our score — by company, across industries, and overall — what makes it worth having is that I could take it with me and share it with others.

Say I’ve flown Delta for a decade. Their people don’t mind seeing me coming, and the record shows I hold up my end by showing up for the flights I book. Tomorrow I’m booking a flight with United for the first time. No status, no history, nothing on file to give them a sense of how good a customer I may be for them.

I should be able to bring some of my history with Delta over to a net-new interaction with United. Not all of it, and not automatically. In this system, I would choose what I share with United. For example, I could share:

1. Proof from one company — like the trust I’ve built with Delta specifically

2. A score for a whole industry — like my score across all travel industry businesses I work with

3. One overall number — to give a bigger picture

In this scenario, the UI and interaction will be simple, while the specification limits what United would see.

Smooth and simple for everyone involved. United knows it has a potentially great customer coming its way, and I would hope to get some extra treatment because of that.

What’s next from here

There are still plenty of questions — and concerns — around a system like this. The four problems I focused on while developing this:

1. Personal reputation measurement systems have been tried. Peeple, Traity, Karma, Deemly. Klout turned influence into a public number and shut down in 2018.

2. The law limits what a score like this can do to people, and it should.

3. Human judgment carries bias. There’s a lot of research showing that rating systems inherit bias in profound and non-obvious ways.

4. Any measurement can mistake a bad moment for an entire person’s complete package.

And then, of course, there are the science fiction-fueled worries from anyone who’s seen “Nosedive,” the Black Mirror episode about a five-star society.

All of those concerns and questions shaped how Conduct was developed. The reasoning behind each decision and the current status of the whole thing have been published. There are also more details on the twelve rounds of testing I did, where I brought in adversarial AIs (Claude, ChatGPT, and Gemini) to give each of them invented scenarios and had them rip Conduct apart.

That has helped make it stronger (while it killed off several ideas I liked: simple averaging, automatic portability, blockchain, and more).

The work so far has shown the potential for Conduct. I believe it could fulfill CQ’s mission to make things better for everyone. By having business records sitting alongside human reviews, by mixing facts with feelings, by sharing your CQ only when you want to, this could take off.

But it needs work. It needs other perspectives. It needs to see whether this resonates with anyone other than me and the handful of people who have tolerated the incessant nerdfests around all of this.

Which is why all of the work is now in the open.

Conduct is the site for us, as consumers. It shows what a CQ could mean for you, what it protects, and what it refuses to do.

Open Conduct is the standard. The spec, the protections, the open questions, the record of what got decided and why. It’s version 0.5.

There’s still time to shape what this becomes, and I want other people’s fingerprints all over it.

Good conduct should count

Here is where we land this exploration. And it goes beyond the original catalyst of Uber ratings.

There are two reasons I think this matters now, and not someday.

The first is that brands already carry a lot of who we are. People build pieces of their identity around the companies they choose, and those relationships mean more than a transaction.

The second is that agents are starting to handle the transactional work for us. Booking, comparing, buying, negotiating. As more of that moves off our plates, the moments where a person deals with a person get rarer, and what happens inside them counts for more.

Human attention will become one of the best things a company can offer.

So.

Maybe this proposal catches on. That would be magical.
Maybe someone reads it and builds something better. That would be awesome.
Maybe it eventually extends to other aspects of human-to-human work, such as quantifying relationships among clients, partners, and teammates.

Wherever this goes, the aspiration is simple:
I want businesses to notice the people who make human work better, and I want those people to find each other more easily.

With more of life running through screens and humanity becoming more disconnected, selfish, and rude, the people who invest in being good people are the ones worth noticing.

Okay, we’ve made it to the end. Now the biggest questions are:
Would you want to know your CQ? And what do you think of all this?