Trust, Business, and Game Theory
Trust is the foundation of every business, but it’s often taken for granted. I buy from Amazon because I trust the delivery time, convenience, and quality of their service. I use Google to search the internet because I trust that it provides me with the most relevant results. I don’t actively think about the underlying trust when using these services— I expect it. The second that I discover something better, I’m switching. It’s as simple as that.
Trust is a great example of Hume’s Problem. Each time I have a great experience with a service, I trust it a little bit more. It also means that trust can be subject to Black Swan events. The second I lose this trust, I leave.
So what? I’m just one person. That won’t bring down a business.
Many people don’t realize it, but every industry has it’s own watercooler. Techmeme is the classic (and most familiar) example of the watercooler of tech. What happens when one really bad experience finds it’s way to the watercooler?
You might say that this example isn’t relavent to your industry, but it is, it’s very relevant. It’s the perfect example of what not to do.
I can only say that I’ve watched one business squander a $1 billion opportunity due to a loss of trust (ok the mortgage industry did that too… so 2 businesses). Not only did it miss this opportunity— it has completely collapsed.
The last industry you would expect to have a watercooler (but it exists and is thriving) is online poker. Two Plus Two Publishing is home to the most popular poker forum on the internet. Opponents (who compete for money on a regular basis) openly discuss hands in an attempt to improve.
18 months ago, Party Poker withdrew from the U.S. online poker industry, giving three smaller sites, Full Tilt Poker, Poker Stars, and UltimateBet an opportunity to capture bigger market share as players flocked to their sites. Shortly thereafter, a cheating scandal erupted where several players lost a significant amount of money (Taylor lost $25,000, and others lost as much as $150,000). That’s a lot of money and UltimateBet has been blasted by the watercooler. UltimateBet has effectively lost its entire userbase as a result of this event. Sure, there have been other cheating scandals. There’s big money at stake. However, none were as poorly handled as this.
I have a feeling that most people will say that this is irrelevant, and people only left because there is money involved. People might also claim that this happens on a daily basis and it only received attention because it happened to the best players (at the highest stakes). I don’t disagree with the latter, only the former. You can compare this cheating to being scammed on ebay, spammed on facebook, twitter crashing, etc, because its something that you care about, relative to your respective services. There is big money at stake in every business, but on the web it often comes in the form of data. As soon as people fear that their data is being compromised (many people are weary of Google), they leave. But what about this notion of a few people’s black swan ruining it for everyone else? Again, this is an important trend— it happens everywhere. People like to follow the best, and you can often see the same tech writers on Techmeme every day. Ideas and opinions often diffuse from these watercoolers— and they can help or harm a business.
Let’s oversimplify trust into the classic game of cheap talk. The user has two options: don’t speak or speak. The company has two options, act or don’t act. These yield the payoffs of:
Don’t Speak, Don’t Act: 1,1
Don’t Speak, Act: 0,0
Speak, Don’t Act: 0,0
Speak, Act: 2,2
Because trust is an example of Hume’s problem, it is also relative. I may trust a service but the service may know something I don’t. If that’s the case, it is better for a company to be forthcoming, which actually re-establishes trust. This is why trust is actually a signaling game— again, I’m simplifying.
Obviously, it is better for companies to listen, and act (don’t speak, don’t act would indicate no problems). UltimateBet failed to react (18 months later!) which is why it’s users left (thus, no gain for anyone). Time is important. If you’re at a restaurant and your meal sucks, you don’t want a free one 18 months from now! You want it immediately (or a coupon for the near future).
Twitter is another great example. Mike Arrington complains, and they listen. They’ll fix the service (what he wants) because it’s in everyones best interest. Problem solved. Stop complaining, Mike.
Trust is what makes anti-stealth so interesting. You avoid the hype black swan by being honest, open, and trustworthy.
It’s clear that trust is important and there are still many unanswered questions, but I’ll end this with the big one: what type of event causes a user to not trust a company like Facebook or Last.fm?
Notes
-
quangtran reblogged this from jeremystein
-
jeremystein posted this
