Bitcoin for B2C Payments, Part 2: Merchant Savings

| Originally published on on February 7, 2014 |

In my previous post on payments, I went over the different factors that would influence fees in a Bitcoin-based payment system. In this post I’ll try to be more concrete about potential merchant savings.

First three notes:

  • Let me emphasize again that any numbers discussed here are just rough estimates. The card system is notoriously complicated and fees vary from transaction to transaction based on a large number of variables. To get an idea, take a look at the the list of different interchange fees (PDF) in the Visa network. 
  • The numbers in the post below assume a 2% transaction fee but the average fee is likely higher (especially for smaller merchants).  
  • For this discussion I’m going to ignore any long-term (more than 30 days) credit given by banks to consumers. This shouldn’t distort the fee discussion much since we can assume that banks cover the cost of long-term credit via the interest they charge on balances, not via the interchange fees.

So where could we reduce fees and by how much?

1) Interchange is the largest target (~1.5%), paying the issuing bank for:

  • Rewards – in the form of cash, miles or “points” given back to consumers (~1%)
  • Short-term credit – allowing consumers to pay for their purchases up to 30 days later and assuming the resulting risk
  • Customer service – any necessary interactions to answer questions/resolve issues
  • Convenience – ability to use a card vs. cash, itemized purchase list at year-end

To make things easy, let’s assume the following:

  • Consumers are ok with paying for everything right away (no short-term credit) and since Bitcoin is a “push” payment rather than a “pull” payment, the risk of consumer default disappears (see this post by Richard Brown for an interesting discussion of push vs. pull payments).
  • Universal access to a very easy-to-use and functional mobile Bitcoin wallet (this is a big assumption).

This eliminates all costs except rewards. Could other benefits (to be discussed in a subsequent post) convince consumers to forego some rewards in the switch from cards to Bitcoin payments? Perhaps. Let’s say that rewards can be cut in half, to 0.5% of transactions.

That gives us possible savings of ~1% on interchange.

2) MSP/ISOs get a small piece of the fee pie (~0.1%), but this is low margin revenue because they have to sell and service the merchants and they often see high customer churn. While the current players may get replaced or their role may evolve, someone still has to do this work. So for this analysis I’m going to assume that this fee won’t go away.

No savings here.

3) Processors probably make around 0.15% and in a Bitcoin world they are entirely replaced by miners. However, before we count this as pure savings, we should consider the transaction fees that are given to miners in the Bitcoin system. Today they play a small role in giving miners incentives to process transactions (incentives mainly come from new Bitcoins), but their role will increase over time. But with that caveat, let’s assume we can save the full processing fees in the short/medium term.

Possible savings of ~0.15%.

4) Card associations (e.g., Visa, Mastercard) are very high margin businesses (see Visa’s 40%+ profit margin) since they coordinate the whole system but don’t do much of the work. In a centralized world, this is a high-value position. In a de-centralized world, it’s unnecessary.

So would Bitcoin eliminate their take of ~0.15%? It’s possible. But there is an interesting thing to note: a profit maximizing business that serves both merchants and consumers (e.g., Coinbase) will be very tempted to layer services (and their brand) on top of the Bitcoin system and extract additional value. Their ability to do so will depend on competition and other market dynamics. But for now let’s assume the full amount will be saved.

Possible savings of ~0.15%.

 5) Acquiring banks’ fee of ~0.15% pays them for doing primarily two things:

  • Receiving funds and making them available to the merchant.
  • Assuming the risk of the merchant going out of business or being unable to honor refunds and chargebacks.

We can assume the more streamlined Bitcoin system will eliminate the need for the first activity, but the risk is not going away. In fact, in a push payment scheme where consumers no longer have easy chargeback recourse, it will be important for someone to assume merchant risk. So let’s be aggressive and say that half the fee can be eliminated.

Possible savings of ~0.075%.

So where does this leave us? Saving 1.375% of the initially assumed 2% in fees – a fee reduction of ~70%. That’s actually pretty big, but it’s a very aggressive estimate. Specifically, the biggest areas where it may be wrong are (in order of likely impact):

  • Assuming consumers will be ok with only 0.5% in rewards
  • Assuming no transaction fees paid to miners
  • Assuming no fees are extracted by a “network brand” 

But this isn’t the whole story. Let’s not forget that today (and for the foreseeable future) merchants will need to exchange most, if not all, of their Bitcoins to local currency in order to run their business and to avoid exchange risk. At today’s going exchange fees, this adds ~1% right back in.

So that leaves us with savings of 0.375% of the initial 2%, or roughly a 20% reduction. Still decent, but not a huge slam dunk.

But that still isn’t the whole story. What’s left to consider?

Reduced transaction fraud. Today, when credit card fraud happens, it is most often the merchant that is on the hook – not only do they lose the money that was paid for the fraudulent purchase but they also have to pay a chargeback fee. With Bitcoin payments, this fraud will be much less common. This is because:

  • Payment info (e.g., card numbers) will not be handed out freely (e.g., to a waiter) or transmitted over open lines (e.g., unencrypted Internet connection)
  • Payment info will not be stored in hackable merchant servers

This means that merchants will have additional savings. These will be larger for online merchants (as “card not present” fraud is more common) but may still be significant for bricks and mortar merchants. Can we quantify this? Well, in a recent panel (from the New York State Department of Financial Services Bitcoin hearings – see Panel 2 Day 2), John Johnson of gives out these numbers:

  • 14% of their orders are screened because they raise possible fraud flags
  • 2% of orders are canceled because they are likely fraudulent
  • 0.18% of sales still come back as chargebacks

Estimating how these numbers would change for the average merchant is too speculative, even for this blog post. But reducing the necessary screening costs and chargebacks would result in real savings, especially for merchants that are less sophisticated than Overstock.

One final note on this aspect of the equation: the card networks are pushing US merchants and issuing banks to fully migrate to EMV (chip) cards by October 2015. In other countries this move has resulted in reduced fraud and that is likely to happen in the US as well, making the fraud advantages of Bitcoin smaller. But it’s worth noting that while EMV will reduce some types of fraud, it won’t reduce the risk of a retailer’s servers getting hacked.

Based on all of the above, we see that merchants could reduce their transaction fees by more than 20%, perhaps much more in the longer term if exchanging out to local currency becomes less necessary. As discussed in the previous post, saying Bitcoin will reduce fees to 0% is not reasonable – but the savings we’ve discussed in this post are very significant and worth exploring for merchants.

In the next post on this topic we’ll talk about the pros and cons for consumers, what could encourage them to make the move to pay in Bitcoin and we’ll then bring everything together.