A couple of weeks ago, I was introduced to the bitcoin-based tipping service for Twitter, ChangeTip, by Leslie John Dilley. It is a fascinating thing to experiment with, and you should try it out if you’re interested in bitcoin. What makes it particularly interesting is that you can define your own pseudo-currency units called “monikers”. The result is a deceptively simple-seeming UX that not only lowers online tipping friction dramatically, but nails the psychology of tipping in a very powerful way. You simply mention who you want to tip in a tweet, and how much (using either a standard unit or a personalized moniker), and cc @ChangeTip. The service then sends out a tweet inviting the recepient to collect. The second tweet looks like this (notice how the 2nd and 3rd examples in the image are using personalized monikers; “bits” in the first one is one of the standard system-wide monikers):
Besides making you feel all important because you get to name your own currency, it is interesting because it helps you shape the social perceptions that ride along with a tip. One of my personal monikers, for instance, is “refactorings” where a refactoring is worth one penny. If I tip somebody 100 refactorings on Twitter, it means they get a dollar if they choose to collect, and if they know me, they also know that they’re getting it because one of their tweets helped me see something in a new way.
The idea of shaping the perceived meaning of a transaction is a hugely important one I think, and opens up very interesting new territories for economics. One of its effects might well be to increase the importance of tipping and decrease the importance of “meaningless” basal transactions. This is an interesting development because one of my assumptions so far has been that digital transactions break tipping cultures.
I have now changed my mind to the polar opposite view: going digital will eventually strengthen tipping culture to the point where the tipping economy might even become bigger than the basal transactional economy.
The reason we’re not yet seeing this happen is because the old economy understands tipping in a far weaker way, which does break when moved online. An example is Starbucks. Here’s how their app does tipping.
After years of not supporting tipping, they caved and allowed cash tipping at the counter. A little later, they caved further and the Starbucks app began supporting tipping (I can’t recall when). But the design is so broken — by intention I suspect — that you have to conclude they don’t wan’t you to tip. I am a case in point. I am generally a good tipper at coffee shops where I pay cash, but not at Starbucks or other places that offer pre-loaded card or app schemes. I like cashless transactions so I adopt such things instantly. But the tipping experience invariably breaks.
I am not even a UX guy, and even I can think of 10 ways to make the Starbucks design way more natural so I tip at the same or higher levels than at indie cafes.
Here’s the problem with the app in brief. Notice how the tipping is not part of the payment workflow at all. In the first version, the tipping option was so buried, I didn’t even spot it when it was first introduced. In this version, it’s more visible, but you have to do it separately, within two hours of the transaction, but as a separate action. So they want you to move away from the counter in space and time and do it later. But I don’t think this is a perverse dark pattern. They have legitimate reasons to introduce friction in the tipping. Maybe they are even rationalizing it as a way to make lines move faster.
My claim: this is an example demonstrating that porting cash-based tipping cultures to the world of digital transactions — cash-free and wallet-free — breaks tipping. But doing tipping right in a digital-native way can have the opposite effect: making tipping culture far bigger and much more complex than it is today. The key to this possibility is that done right, the perceptions that ride along with tips can be programmed by you or the receiver in a dynamic way, instead of being a static function of the prevailing normative tipping culture.
Cash Tipping Culture
Now, I am going out on a limb here, but I’ll speculate that I am not the exception. I suspect very few people use the tipping mechanism in the Starbucks app, relative to tipping frequency/amount in cash transactions.
The reason is obvious: most of the time, you want tips noticed and perhaps acknowledged with a thank-you or a nod. This is why you want it to be as naturally and frictionlessly integrated into the payment experience as possible. Tipping is not an anonymous-charity kind of transaction for most of us. It is a transaction that serves multiple intertwined purposes: goodwill, gratitude for exceptional service, relationship maintenance, and occasionally, charity/gift motives.
The purposes of tipping are much more strongly shaped by the context than the giver/receiver perceptions and intentions. Individual intentions aside, in the United States, there are six tipping contexts in a hard-edged economic sense.
- Where tipping is factored into the microeconomics, like a restaurant that underwrites a minimum wage but expects waitstaff to make the bulk of their income on tips. Here you want to tip for the same reason you want to pay at a grocery store — not doing so is a kind of theft of somebody’s income under that social arrangement.
- Where tipping is expected, like a restaurant, but not as strongly factored into the economics (it creates a bonus rather than a living wage, say), you want to make sure you’re seen to tip so you aren’t thought of as an asshole. Not tipping here is a kind of free-riding on the tipping of others.
- Where tipping is in a sort of gray zone of expectation, like in coffee shops with a counter-jar. If you’re a regular, you want to tip to reinforce a positive relationship (and prevent retaliatory bad service on future visits), but it’s not completely expected.
- Where tipping is not expected, and people are surprised if you do tip, but will not refuse.
- Where tipping is not expected, and will be refused if you try to offer.
- Where tipping is not expected, and will not only be refused, but seen as an offense if you try to offer (perceived as insulting or assuming bribe-ability for instance). Possibly even a criminal offense.
Within these six levels, there is vastly more variety. Each tipping context is in fact what I like to call a tipping microculture.
This is complicated enough that there is a whole lifestyle blogging cottage industry devoted to figuring out if, and how much, you should tip in different situations. There’s probably enough variety here for a dozen PhDs in anthropology.
Across these contexts, I’ve encountered two ways people think about tipping: pragmatic and tribal.
Pragmatic tippers explain the psychology of tipping in unsentimental ways. Mine, for instance, is “the rich tip to bribe, the middle class tip out of guilt, the poor tip out of a sense of solidarity.”
When I tweeted that, a few people were annoyed. Those are tribal tippers.
Tribal tippers explain the psychology of tipping in sentimental ways, in terms of romanticized relationships (these are the sorts of people who value “small, local coffee shops where you can really get to know the owners and there is a sense of community”). To tribal tippers, all tipping is about contributing to a community, reinforcing a pattern of relationships, expressing sincerely felt gratitude and so on. It’s priceless economics rather than priced economics (though tipping behavior does not always map to general guardian versus commerce personality type).
Whatever your preferred scheme of explanation, it is clear that we generally want to load tipping with cultural meaning of some sort. Tipping is smart money, where the “smarts” is in the form of social and cultural capital rather than intelligence. Maybe we should call it political money or sentimental money, depending on whether the giver is a pragmatist or tribalist.
The meaning is about shaping the power relations for pragmatists (reinforcing a pragmatic status relationship and maintaining it in a healthy state) or community dynamics for tribalists (elevating a stranger to an in-group friend of sorts via a shared priceless value).
The Starbucks app breaks tipping culture for both tribalists and pragmatists, because it takes a small-business model tipping context (type 3) and turns it into a big business model context (which is generally either type 4 or type 5, and sometimes even type 6).
In other words, it isn’t the Starbucks app that breaks tipping culture. It’s the nature of large corporations.
Corporations Break Tipping
Transactions enact two relationships: between business and customer on the one hand and the service-provider and customer on the other.
Tipping is an add-on voluntary transaction that not only shapes both relationships at once, but also a third, generally hidden relationship: between business and service-provider. The tension that emerges with corporations is: who gets to shape how much of the overall transaction meaning, and in what way?
For a concrete example, consider a $20 check and tips of $1, $3.50, $4 and $6. In the US that would be “stiffed”, “miser”, “good”, “very good.” The tip amount from $1 to $6 will shape the perceptions for the base amount of $20 as well, for both parties.
Who feels what towards whom after each of those transactions, in each of these triangular situations?
- When the business and the service provider are the same, such as in the case of a family coffee shop, the person you relate to and the entity you are transacting with are the same.
- When the business and service provider are personally related, such as a restaurant where the waitstaff can talk to the owner about tipping policy. Here your main transaction is with one entity, and your tipping transaction is with another, but both are locally situated individuals, strongly related and can negotiate things. The owners may also participate in service occasionally.
- When the main transaction is with a distant, impersonal corporation and the tipping transaction is with a person, you have messiness.
The first case is fairly trivial. Depending on the type of business, and whether the owner identifies more as a pragmatist or tribalist, they’ll react appropriately. The Soup Nazi will kick you out if he doesn’t like you, no matter how much you tip. A more pragmatic owner-server will be nice to all, no matter what the tipping level.
The second case is harder. Interesting things happen when owners participate in employee work. There was a Gordon Ramsey show for instance, where the owner of a hotel would take a cut of the tips to the waitstaff on days he helped out with waiting tables. That was (correctly) viewed by all but the owner as a douchebag move. Some people decide whether to tip or not based on whether or not the person serving them is the owner (no tips for owners, tips for non-owners).
The third case is the messy one. There are two sub-cases, low-cost commodity leader and branded.
In a low-cost-leader business, the messiness can be exploited to favor the corporation. For example, a corporation can pay very low baseline wages, and put the moral burden of ensuring a good income for its front-line service staff directly on customers. This is not generally a great idea for a strong brand, because there is too much variance in service quality driven by tipping.
It works for low-end, low-price businesses where the business only wants an unsentimental “we are the cheapest providers of this commodity” relationship with you, and either ban tipping to make the whole relationship unsentimental, or leave the rest of the relationship potential to the front-line staff to work in their favor or not, leaving sentiment as a free variable.
For a branded, premium business like Starbucks on the other hand, this breaks down. This is because branded businesses believe they need to use authoritarian command-and-control models to control the customer’s relationship to the brand, otherwise Horrible Things Will Happen.
This leads them to view phenomena like tipping as subversive and disruptive threats to consistent quality of service. This then, leads to apps like the Starbucks app, that cripple the tipping UX and prevent relationships from forming.
This fear is completely ungrounded, but is nevertheless the assumed reality for most brand-driven businesses.
Having the Brand Cake and Devolving Autonomy Too
Starbucks very badly wants to own the entire relationship with you the customer. I know this is true because I recently took a very long survey where they were basically strip-mining my brain for everything I could possibly tell them about what I wanted in a future loyalty program (I’ve been a Gold member for years). The assumption of the survey was clear: it was entirely about me and Starbucks the company. The baristas were interchangeable cogs in the picture.
For a business like Starbucks, which seemingly believes (incorrectly in my opinion) that it needs to shape my relationship to the brand with authoritarian comprehensiveness, tipping is a quality-assurance nightmare. A much more natural tipping experience would basically make barista behavior too strong a function of tipping to be “managed” effectively under their current management model (which we caught glimpses of around the time of the free college stunt).
With the current app, I don’t even know if the barista who served me knows that I tipped, let alone acknowledging it. By separating out and moving the tipping transaction far away in space and time from the main transaction, the app retains the mechanical possibility of tipping, but kills its relationship development potential.
You could easily imagine a very different app, where you had the option to tip pop up right after the barcode scan, and if you did, the barista in question would have an opportunity to than you right then, or “Like” your tip later, the way you Like posts on Facebook, and maybe even leave a comment. Barista and you would learn each other’s names over time, you’d beat the awkwardness of cash tipping (where if the barista happens to look away just when you drop a buck into the jar, you don’t get “relationship credit” in whatever form you perceive it), and create a smoother relationship.
You could even create an entire temporary mini-social-network involving you, the barista who took the order, the one who made the drink and anyone you said hello to. If you’re a regular, they could all wish you happy birthday when they made your free birthday drink. Maybe friendly baristas from other stores could wish you too if you bought a birthday drink at one store. Individual stores would have unique, signature microcultures without breaking the uniformity and consistency of experience that cloud mice like me prefer (and the reason we prefer Starbucks over indie coffee shops).
For customers who don’t want such intimacy, there would be an option to retreat to an anonymous tipping role, where it would not be clear whether they tipped or not. Kinda like reverting to the “Classic UX” in software applications.
But a nightmare for Starbucks under its current management culture (at least as it appears from the outside).
If I were Starbucks CEO, this would seem to me like opening up a huge can of worms. Creepy guys could stalk pretty baristas through such a mini-social network. Baristas could grow to like/dislike particular regular customers enough that it starts affecting quality of service, all the way down to that business-killing customer suspicion, “did she spit in my drink because I didn’t tip?”.
With such service quality variance, brand coherence would weaken, and corporate marketing would lose control of the brand narrative. Within two years, Starbucks would have a brand perception on par with BP after a major stalking scandal.
These are all completely imaginary fears.
There are ways to shape corporate culture and evolve the design in a way that keeps out bad behavior, creates a positive kind of variance in quality. For example, think about Zappos’ “extras” philosophy of selling. Baseline service is great, and sales agents can do more to surprise/delight customers (like throw in free faster shipping or a card). It turns the weakening of brand coherence into a strength. If you shop at Zappos, you can expect to be pleasantly surprised frequently, above and beyond a baseline consistent experience. You don’t have deterministic expectations of quality service, you have probabilistic expectations that are skewed positive, and bounded below.
Google does this in some minor ways on the marketing front (by turning the search engine logo into a varying, morphing entity that does PR for the company, instead of being rigidly dictatorial about it), but doing it on the sales front in a retail, face-to-face business, is much harder.
But not doing it puts you on the path to commoditization through robotic automation. Starbucks is now introducing pre-ordering on the app so you can get rid of that counter-friction and simply pick up your drink. It’s only a few steps from there to a business where robotic coffee machines make the drinks and there are no baristas. The grilled cheese chain, The Melt, is already far along that road.
I wouldn’t mind that kind of coffee shop, where you’re essentially inside a very large and sophisticated coffee vending machine. Premium coffee would certainly be half as costly.
While I am generally all in favor of automation, in this case, I think it would be short-sighted for premium coffee shop chains to pursue that path. The business of robotic cafes will be huge in the future, but the business of non-robotic cafes will likely be even more huge. Because coffee shops for most of us, even pragmatists, are relationship spaces. Spaces where part of the value is regular low-intensity ritual interactions with familiar faces. Without that, I’d simply fill my mug from my home coffee machine and go work in some public plaza. There would be no real reason to even go to a cafe.
Shaping a powerful tipping culture where the variance in experience is positive and the source of the cafe ambiance is not an easy problem, but somebody will likely crack it. The result will be the first coffee shop chain where the global brand is associated with certain assurances of quality and consistency, and each individual store also has a unique tipping microculture.
None of this is rocket science by the way. It’s a well-understood, but not widely-known management philosophy that I talk about every time I blog about business stuff: John Boyd’s philosophy of warfare applied to business, as compactly laid out in Chet Richards’ Certain to Win.
The basics are simple enough to compress into a single sentence. CEO sets clear intent (Schwerpunkt); HR develops strong trust culture (Einheit); operations focuses on developing strong, instinctive skills culture through tacit learning (Fingerspitzengehful); everybody manages/is managed through a cascade of mutually negotiated “contracts” that devolve as much autonomy as possible to lower layers (Auftragstaktik); the business relies on loose and agile coordination rather than tight synchronization/command-and-control.
Unless you’ve been living under a rock for a decade, none of that should be news to you. Pull that together in a coffee shop chain and you’d get Blitzkrieg customer service.
It’s basically the idea of fourth-generation warfare — creating fear, uncertainty and doubt (FUD) by getting into the opponent’s decision cycle — turned on its head. Same mechanisms, but used to create delight and a sense of serendipity in the customer experience rather than FUD in the “enemy experience”. Like being in Disneyland.
I don’t expect this to happen at Starbucks, because despite their use of a lot of digital technology, they’re fundamentally an old-school company that does not instinctively get this stuff by way of corporate habits. But maybe one of the younger, smaller, upstart chains will get this and run with it.
This is just the tip of the iceberg of what’s possible with imaginative use of tipping microcultures.
Dynamic Tipping Cultures
Even if executed in the most enlightened traditional corporation, the tipping culture context of an old-school business is going to be constrained. Tipping microcultures at individual locations are going to be pretty uniform due to command-and-control being the majority approach to scaled businesses.
On the toxic end of the spectrum, you get tipping as a part of rather ugly customer-guilt-as-employee-income-security HR strategy. On the positive end, you get Blitzkrieg-like delightful customer service in a situation that keeps business, service provider and customer all happy in a win-win way (no company does this perfectly today, but Zappos comes closest, though I have other criticisms of how they work). The least privileged of the three, the service provider, enjoys much greater autonomy and optionality (opportunities to make way more money through creativity than in even a very high fixed-pay culture).
But in all the cases, it is a case of fairly fixed roles, a fixed range of cultural meanings that can be created through transactions, and local tipping microcultures that are inherited unchanged from global corporate culture.
When you go digital-native, the options start to broaden. In the rideshare industry, Uber does not do tips, while Lyft is essentially pure tip in a legal sense. As far as I can tell, neither company is really exploring tipping as a powerful relationship shaping variable, because both are engaged in a pure low-cost-leadership race with each other and the taxicab industry. The industry defaults are not good: taxicab tipping culture is not the pleasantest thing out there.
But once this cost-driven phase levels off and there is room for differentiated, branded experiences (I am told some Lyft drivers tell jokes or do magic tricks), how tipping culture is shaped may make or break rideshare companies.
The most interesting case, however, is when there is no corporation in the picture at all, and “tips” are being exchanged in a network economy business where the base product is “free” in old economy terms. The world of ChangeTip in short.
To some, such an idea sounds horrible and even insulting. David Pogue for instance, while ranting about writers being expected to work for free, argued that expecting writers to work for tips and sponsorships while providing content for free is like expecting programmers to work for free mousepads.
My point is, if the mousepad is valuable enough and big enough, this can actually be a sustainable scheme. Despite what skeptics seem to think, there’s no natural law dictating that tipping has to be on the order of 20% of billed value.
If you’ve been a long-time reader, you may recall back when I had the “buy me a beer” plugin installed (with coffee instead of beer). The author of that plugin claimed (very credibly) that the UX tripled tips over a “leave a tip” UX. My experience validated that claim in spades.
Why? Because it reshaped the power dynamics. We normally associate tipping with a high-to-low status service context (“here, lowly barista, I, grand-and-always-right customer deign to leave you an extra buck”). “Buy me a beer” recoded that default meaning to “nice post buddy, lemme buy you a drink for that.”
I discontinued that plugin because keeping track of hundreds of $3 “coffees” was a book-keeping pain, not because the idea isn’t sound.
What ChangeTip does is completely generalize that logic with its system of monikers. Now you can define a set of monikers to tip people while recoding the meaning of the transaction in completely arbitrary ways, ranging from, “You are a God, here’s a tribute” (suggested moniker: 1 tribute = $10) to “You are a desperate Twitter armchair philosopher, squirm for this measly tip, you worm!” (suggested moniker: 1 grovel = 1 penny).
You can imagine other possibilities. Imagine for instance, dining at a restaurant where the meal is free, but there is a normative culture of tipping where the amount can have very varied connotations set by the receiver. If I leave a tip of $1 on a meal clearly worth about $50, the waiter can give me a final receipt where the $1 is coded as “100 douchebags.”
If ideas like ChangeTip catch on and are refined, I am willing to bet that the size of the tipping economy will grow by several orders of magnitude, with much of the transactional volume happening in tip-only network economic contexts where the base service is free.
And I am willing to bet that instead of a single universal kind of gift-economy currency, like Cory Doctorow’s somewhat collectivist idea of whuffie, it will be a huge universe of monikers and meaning-tags that effectively make every transaction its own currency. Tipping microcultures will fragment further into tipping nanocultures. Cultural economy diversification, not homogenization.