Pricing Research: the power of good questions
Twitter is currently on fire and Elon’s trying to figure out how to make money. We’ve heard engineering commentary about how shipping fast is going to be the end game for that organization. He’s decided to charge people $8 for a blue check mark and some additional perks. He’s trying to make sure all that money he invested and the debt he incurred doesn’t sink him. So he’s decided to change pricing. My guess is it won’t bring in the revenue he’s hoping for, and the reason is that he’s not doing very good pricing research.
There are some very logical pieces to pricing: gate a necessary feature, something people absolutely need, and they’ll figure out a way to pay for it.But pricing is inherently psychological, not logical. A common way to use psychology to price goods is to tier them. You have a low, mid, and high tier. People will, on average, choose the middle tier when presented with three options. Twitter is built on consumption. If a product is “free” then you are the product. And that’s mainly the case for Twitter right now. So consumption is what Twitter needs to drive. Monthly subscriptions, when done well, actually do increase consumption.
Talk to people, but ask better questions
The first step in pricing research is to actually talk to people, to your representative user base. Kind of like this:
Oh good, you’re thinking! I can just do a Twitter poll and my questions will be answered and it’ll all be fine. No.So here’s what this poll gets right. It’s talking to the relevant user base (Twitter users), about a feature or change that will impact them.
But this is where it all goes sideways. People looooove asking what someone would do in a given situation. “How much would you pay for X”, “If we made Y, would you buy it”. I cannot believe that we are here in the year of our lord two thousand and twenty-two, still asking this same question. It is enraging.Stop asking people what they WOULD do. People don’t know. We all love this idea that we know ourselves. That we are logical beings making logical choices. And it is a lie.
Additionally, if people know that a pricing decision will impact them directly, it’s in their best interest to answer strategically. So it’s not a good indicator of what people will actually pay, because people are answering in their own self interest.
So. Stop asking future based decisions.
Instead, ask what people currently DO pay for a similar service. A better question would be:“What monthly subscription services do you currently pay for?” And then, if you’re gonna get real spicy, you ask the mother of all follow up questions. “Tell me more about that”.
Because people have absolutely no clue what they would do in a given situation. But they’re really great about telling you what they’re already doing. It’s up to you, the researcher, to analyze that information and find the pattern. Find the common thread. Talk to your users.As you can see here, Elon started a conversation with Stephen King about pricing models. I knew we were walking around in growing dystopian hellscape, but this is quite a funny little dark comedy amidst the flames.
So, first of all, hooray! You’re talking to a user! That’s step 1. But here’s where this breaks down.
“We need to pay the bills somehow! Twitter cannot rely entirely on advertisers.”Cool, cool. We get it. We all have mortgages, and private jets, and submarine cars to pay for. We totally understand.
But here’s the thing. Your users, the people on the platform, don’t care about your bills. They care about their experience. So how is your change going to make Stephen King’s life better?
And then we step into this question again: “How about $8?” Stop asking this. People don’t know. They think they know, but they don’t.
A better strategy for deciding pricing
When I’ve done pricing research in the past, we always did real world testing. We ran what’s called a multi-variate test and created three separate pricing tiers. The tiers were determined by analyzing how much competitors charged, and then pricing ourselves slightly higher than that, with 3 different options.
We pulled prospective customers, and presented them with the pricing when they got to the end of their journey. And then the person would have to use their own money to pay. People are far more conservative when they’re using money from their own wallet. While this method was by far from perfect, it gave us a much better signal around what people were willing to pay.
An even better method is the BDM (Becker-DeGroot-Marschak) method. A person says the maximum they’re will to pay for something. Then, they draw a random number out of a lottery. If they number they draw is less than or equal to the price the maximum price they were willing to pay, then the person pays for the item and receives it. If the drawn price is more than the amount the participant said they were willing to pay, then the individual pays nothing and receives nothing.
Participants are then incentivized to be truthful about how much they are willing to pay, because there’s a chance they won’t receive the product if they give a price that’s too low. But, they have to spend their own money to get the product, so the price they give also reflects how much they’re actually willing to part with. For this to work properly, the participants need to understand the game, and also actually want the thing that they’re submitting a bid for.
Use this method when you’re trying to understand what people would pay for something in an environment where you can’t talk to more than one person at a time.
Another strategy is the MLP (multiple list pricing) method. Participants are presented with a list of prices. The participants must respond yes or no to each price, yes meaning they’re willing to pay and no meaning they’re not willing to pay. The researcher then randomly selects one price and that price is implemented.
It’s in the respondents best interest to only say prices that are below the price they are willing to pay. Again, participants must understand that the price they pick will be implemented, and it has to be something they want. A list of prices could anchor participants to a specific set of pricing, which is a drawback of this type of study.
**You’ll notice there’s one thing in common with both of these approaches. And it’s this: the person has to want the thing that you’re offering. **Forcing people to pay for something that is optional and people don’t value will degrade trust in the product or service, and that becomes a slow road to decline.
Additionally, adding pricing to certain items, removes the value that people saw in the product to begin with. In the context of Twitter, reframing the pricing strategy as required for high quality verification and maintenance is a much better way to incentivize people to pay than saying they have to pay for verification.
So, what should Twitter do? They first need to understand how their users consume and post on Twitter, and then they need to understand how valuable the platform is to the users. And it is valuable. But maybe not in the way people might think.
Twitter should look at the competitive landscape. Have others successfully charged and what was their model? Who failed to charge effectively and what did the outcome of that failure look like?
If they don’t already know, run a study to understand why people consume Twitter. Who posts and why? Who is only reading and why? What other similar services do these people pay for and why?
Identify the value of the service. In my limited experience on Twitter, the value is the quickness of material from verified sources, ability to find information outside your known social sphere, and the ability to interact with others and receive a pretty great dopamine hit when people like your stuff. Twitter is absolutely built to be addictive; and it works.
Gather two groups. One group of your most active users: people who post a minimum of 5 times per day and have high engagement. A second group of people that are verified and have a high follower count.
Run a BDM study against the value that people find on Twitter. Maybe it’s an ad free experience, maybe it’s access to certain accounts, maybe it’s the ability to edit, or maybe it’s a highly curated experience. The point is, you can get part of the way to an answer on instincts and experience. But what got you here, won’t get you there. To grow and learn you need to actually do research to understand. And in order to understand, you absolutely have to ask good questions.