Over the past year I’ve been asked my thoughts about ‘loyalty points on blockchains’ many times. The thinking seems to be bitcoin -> digital currency -> digital tokens -> loyalty points and at first pass it feels like a natural extension of a theme. People read about cryptocurrency trading and interoperability then think “Wouldn’t it be really cool if I could exchange my loyalty points for other ones, or if I could buy and sell them with real money?”.
This post attempts to describe how I understand the purpose of loyalty points, and in this context, how applicable blockchains are as a technical solution.
First, some terminology:
If two things are fungible, it means that they can be swapped or substituted with no change in usefulness – so I don’t care if I have one or the other. Fungibility is relative and context specific, which means it depends on what you’re doing with the thing.
For example, pound coins are generally fungible. I don’t really have a preference between one specific pound coin and another. Unless of course one is visibly marked and is known to be the proceed of crime; then I would prefer not to have that one. Or if one has got chewing gum on it and I would prefer a fresh shiny one to put in my pocket. So sometimes pound coins are not fungible. Fungibility is contextual.
Pounds and dollars are generally not fungible. Even at today’s fair exchange rate of 1 GBP = 1.40 USD, this does not mean 1 GBP is fungible with 1.40 USD. They are different things and I use them differently.
If two things are interoperable it means that they are capable of interacting with each other. They have some idea of what action to take to make each other do things. Wikipedia has a definition. Interoperable is not the same as interchangeable or exchangeable, because it means ‘systems talking to each other’.
What are loyalty points?
For the purpose of this post I am thinking about airline miles, hotel points, credit card points, supermarket points – ie digital tokens tied to a company or group of companies. Usually I gain points by signing up and buying a product or service. I accumulate points and then I can spend them on products or services, usually with the same company who runs the scheme.
There are single company schemes, for example Tesco Clubcard points, and loyalty schemes participated by a network of companies: popular networks are Nectar in the UK or airline miles eg the Star Alliance scheme.
It gets a little more complex: in Singapore when I use my Citibank-issued credit card in certain shops such as Starbucks, I can get a 1%-2% SGD-denominated “rebate” linked to the credit card that I can redeem in any of those shops. The rebates are linked to my credit card, and are gained and spent by shopping in specific shops, using that credit card. For example if I spend $100 in Starbucks, I earn a rebate of $2 which I can use to reduce my total bill at any shop that is part of the network (as long as I am using that same credit card). I’d call these rebates as opposed to points, because they are denominated in actual money rather than points or miles. Additionally and confusingly, I also get Krisflyer points for Singapore Airlines by using the card which are totally unrelated to the rebates. We love our loyalty points in Singapore.
Loyalty points can have a shelf life. In some schemes, they run out or ‘expire’, ie you lose them if you don’t spend them. This is great for the companies issuing them but not so good for the customer. This puts them in a grey area for fungibility (within one scheme) – yes a point is a point; but a point earned in January which runs out in August is not the same as a point earned in March which runs out in October, and I’d rather have the latter.
The monetary value of loyalty points (to the customer) is usually controlled by the person or company who decides how many points is needed to claim a reward. If I usually need 20,000 points to buy a flight, but now I only need 10,000 points, the value of my points has increased.
Companies love loyalty points
The most obvious benefit to the company is that a loyalty scheme creates customer loyalty, ie customers preferentially choose to repeatedly buy from that company because of the discounts and benefits. This is good for customer retention and stickiness.
Unfortunately there’s a bit of game theory – this only works when other companies aren’t doing it. When many companies have loyalty points, the customer may get ‘loyalty point fatigue’ and shop based on other preferences, cashing in on points as and when they can.
More relevant promotions
Even very basic non-digital schemes (stamp your card every time you buy a coffee, and your 10th is free!) encourage this kind of behaviour. However there is much more value to be gained: by using a loyalty scheme that couples who a person is (along with their demographic data), with what they buy, shops can target customers with much more compelling offers both individually and in aggregate. The classic story from 2012 is how Target found out a teen girl was pregnant before her father did.
Cashflow and profit
Airlines love this: Sell some points in bulk to credit card companies, who distribute them to their customers. This is profitable because the cash received by the airline is greater than the liability created on the balance sheet.
The fair value of a point is should be based on the following:
- The discount that the customer will receive, that is, “the amount the customer will save”
- The expected redemption rate of the points
Both of the above can be changed by the company and allow for management to make a judgment, as these vary widely by company and industry. For more on accounting for loyalty schemes, see IFRIC13; these are international accounting policies that came into effect in July 2008. Here’s a good piece that explains it.
Companies need to manage the liabilities arising from points that have been issued. A liability arises because the company owes someone something in the future: in the case of points, it’s lost revenue opportunity. If there are too many points out there, it increases the liabilities on the balance sheet, which accountants don’t like.
However there are tricks to reduce this liability, one of which is using judgment about what proportion of the points are likely to be actually redeemed. This leads to the next point.
Businesses are the central bank of their loyalty points
Businesses want to have control over the points, and that’s fine. They control:
- The creation of points
- How many points are given out, for what behaviour – are you rewarded just for spending money, or for buying certain products or groups of products?
- Companies can change at will how the points are created and handed out. As an example, see what British Airways did in 2015, angering loyal customers.
- The destruction of points
- How your points will expire if you don’t use them.
- This is quite literally almost the best thing ever for a company, simultaneously booking a profit and reducing a liability, without the provision of the service. Most airlines do this.
- The distribution of points
- How points can be distributed between points owners. Not being allowed to transfer points is preferred by companies, because having many customers with few points each leads to fewer redemptions. Usually some sort of minimum number you need to have before being able to redeem them.
- By allowing points to be transferred or gifted to others, they are likely to migrate from people who won’t spend them, to people who will. This costs the company the cost of the provision of the good or service.
- Increasingly, family schemes have appeared where points can be earned, transferred or spent within a family.
- The value of points
- By controlling what points can be redeemed for, how often, how much, and when, the value of the points is controlled by the company. Airlines change the value points fairly often, for example during sale periods where points can be redeemed to get deeper discounts.
In summary, the companies operating the schemes control the creation, destruction, distribution, and value of points. This is good for business and loss of any point of control would not be desired.
There are examples of having the ability to exchange points, in a controlled manner. For example, in 2014, in an airline industry first, people who were signed up to both KrisFlyer (Singapore Air’s scheme) and Velocity (Virgin Australia’s scheme) could convert points from one scheme to the other at a rate of 1.35 points to 1. This means for every 1.35 points you have of one scheme, you can convert it to 1 point in the other scheme.
For finance people this gives a bid/ask rate of 0.74 / 1.35 (If I want to buy 1 point of the other scheme, it costs me 1.35 points of my scheme, and if I want to sell 1 point of the other scheme, I receive 0.74 points of my scheme). In other words, if I convert 1 point from my scheme into the other, and then back again I end up with 0.55 points. Not very competitive.
This isn’t a ‘free market’ for points, this is a fixed price conversion. Also note that the names and dates of birth on the accounts have to match, so you can only transfer your own points to yourself. This is converting points from one scheme to another, not exchanging points with someone else for something else.
As soon as you can transfer points from one person to the other, they become exchangeable for other points and for cash. A market and market price soon develops. After all a trade is the transfer of asset X from person A to person B, and a corresponding transfer of asset Y from person B to person A.
There doesn’t need to be a formal ‘exchange’ platform, or formal escrow (though being able to escrow the assets helps to reduce the need for trust that the second person will deliver what’s promised).
Without escrow, intermediaries appear who can take the asset from both parties and only pass them to the other party once they have received the assets from both sides. They build a reputation and become the trusted counterparty or intermediary.
So, all it takes for markets to develop, for people to trade points, to buy and sell them with cash or other points, is the ability to transfer points between members of a scheme. The technology to enable this has been around forever, and is already used to distribute points between family members. Reassigning points from one person to another isn’t hard.
The decision to not allow freely transferrable points is not because it’s technically impossible.
When you allow points transfer, you increase the value of points to the customer as they have more utility value, and also increase the proportion of redemptions, because points will have a natural tendency to flow away from those who don’t value them to those who will use them. This has a balance sheet impact because the ‘fair value’ of the points will increase (the redemption rate is higher), and a cost impact because more points will be redeemed instead of sitting idle or expiring.
In summary, with transferability, although companies retain complete control over creation and destruction, they lose control over distribution and to a certain extent value of points.
Blockchains and loyalty points
So, we have seen that loyalty points are not always fungible as they can have expiration dates. We have also seen that companies enjoy the privilege of being the central bank of their own loyalty points, controlling the creation, destruction, distribution, and value of the points. They have decided not to allow transferability not because it’s technically impossible, but because it’s good business.
What are blockchains good for?
Blockchains, as they relate to cryptocurrencies and digital tokens, are good for the transfer of non-expiring fungible digital tokens. (It can be argued that some cryptos with lineage eg Bitcoin are completely non-fungible, as every ‘coin’ has an easily traced lineage, but that’s a post for another time.)
Public blockchains (eg bitcoin and offspring) attempt to solve for anonymity as anyone can create a wallet without self-identifying, and censorship resistance as it’s designed to be hard to block a transfer. They are designed to not allow chargebacks the ability to recall a transaction.
Private blockchains give you the ability to replace the need to trust a centrally-trusted entity by replacing it with technology.
What do companies need?
Companies need a low-cost, secure way of managing their loyalty points programs – the issuance, tracking, and redemption of points. Simplicity (ie a good user interface) for both the customer and for the company running the scheme is key. Security against hacking the database via the front-end or via the back end (or by bribing admin staff).
Low-cost doesn’t mean the cost of selling or transferring points (amending a row in a database is essentially free) – it means the cost of the team needed to run the loyalty operations. If technology isn’t used efficiently, for example if redemption involves two members of staff verifying the wet-ink signature of a form that is mailed in, then this isn’t efficient.
Where is the business mandate?
If technology can make starting and running a loyalty points system cheaper, easier, more secure, and with better visibility over points issuance and redemption, then this is useful. Interoperability with other loyalty programs, and the ability to convert points between programs has been explored pre-blockchain. By making interoperability cheaper between programs, this may have business value. APIs are the key here (a system on a blockchain without API hooks is just as un-interoperable as a non-blockchain system without API hooks).
Business like the control they have over points. Allowing transferability of points reduces that control, irrespective of the technology used to allow it.
Companies don’t want to lose the most valuable asset: data on individuals and their spending habits. They presumably want the ability to censor and unwind transactions should they detect any bad behaviour. So the key defining features of public blockchains (including censorship-resistance) are therefore not useful here.
What about a private blockchain? For a group of companies participating in a loyalty scheme, I can imagine that storing the data on a blockchain on which they all have access may be useful, as the single source of truth about the order of events.
However, I don’t buy the argument that you can do away with the 3rd party entirely. Although the 3rd party may not operate the day-to-day running of the scheme, I believe we will still see 3rd parties whose role will be to coordinate changes.
Security. Perhaps if data were held on a blockchain, held across multiple data centres, this could prevent certain types of hacks where hackers give themselves or their friends loyalty points? And in a joint scheme, by using a private blockchain, members of scheme could know that no other participants to the scheme could unilaterally alter the database. This is a good argument for having a blockchain.
The trust issue goes further than tweaking a transaction or account balance. For example, consumers have to trust that companies won’t devalue the points when you want to spend them. This business decision of how many points a flight costs has nothing to do with how the ownership of points is recorded on a ledger. Given the fact that companies can and do abuse this ‘value of points’ trust like British Airways recently did, blockchains don’t solve for all trust issues.
To conclude, I find little overlap between what companies need as loyalty points solutions and what blockchains are good for. What’s really needed is a review of manual processes involved in administrating loyalty schemes, especially those with multiple participants, and then applying appropriate technology to reduce those costs. A solution should be considered on the usual merits of functionality and total cost of ownership including support and security, rather than on the underlying technology.
If you work on a loyalty scheme, or if you have a thoughts on why a company or group of companies would want to use a blockchain for loyalty points (or otherwise), I’d love to hear from you.