Do remittance startups have fundamentally different cost structure vs. incumbents? What are the primary customer acquisition channels for money transmitters? What can explain massively higher relative valuations of remittance startups vs. established providers? If you are interested in such questions, this article is for YOU.
We will cover the following topics:
Remittances represent around $600 billion in transfer volumes and $40 billion in revenues with 20% going to just two countries and providing a massive lifeline for others:
There is a huge difference in prices of remittances across top outbound countries:
Such data is hard to gather and maintain, so the real prices might be lower:
What is causing South Africa to be 5+ times more expensive than Russia? Most experts claim that it is due to two issues, de-risking by banks and exclusive partnership with retailers by Western Union and MoneyGram:
“A major barrier to reducing remittance costs is de-risking by international banks, when they close the bank accounts of money transfer operators, in order to cope with the high regulatory burden aimed at reducing money laundering and financial crime. This has posed a major challenge to the provision and cost of remittance services to certain regions.”
“… the core issue with WU is their exclusivity clauses that have been used for decades to successfully lock markets to one provider, who can then increase their mark-up fees as there is no alternative.”
Neither is the case with South Africa vs. Russia. It is easy to blame “banks” or “Western Union,” but the real root cause is usually with opaque or corrupt governments that favor banks over MTOs and Fintech.
Let’s first align on which companies are the global leaders in international money transfers for consumers. Some might think it is PayPal or a large bank or some startup, but top providers have been the same for awhile. While banks, as a class, are responsible for majority of remittances volumes (see this SaveOnSend article for details), they can’t match remittance incumbents in geographic coverage:
However, within a particular market, top banks are usually among the largest remittance providers:
With rare exceptions, banks’ customers tend to be infrequent senders who transfer larger amounts (Wells Fargo is a unique story, read more in this SaveOnSend article):
Among remittances specialists, average transfer size is much smaller depending on core customer segments. For CurrencyFair and TransferWise that mostly target expats from developed countries, the average transfer size could be few thousands of dollars. For money transfer companies targeting migrants from developing countries the average transfer size is usually few hundreds. Reasons for why migrants send money also differ greatly by corridor and sub-segments within it. For example, Remitly conducted a survey of its customers in early 2017 to find the following distribution:
- 70% send money to help their families with basic needs including housing, food, and utilities
- 16% send money to cover their personal expenses and investments
- 11% send money to help pay for education costs such as school supplies, books, and tuition
- 3% send money to help pay for medical and emergency expenses
Among remittance startups, TransferWise is a particularly interesting company already surpassing monthly volumes of Ria Money Transfer and of other large money transfer startups combined (see more details on remittance startups in this SaveOnSend article):
Besides these larger providers, there are thousands of other money transmitters specializing in particular corridors and send-receive methods but predominantly via cash agents. Finally, there are dozens of Bitcoin-based remittance startups, but the largest ones have been around for less than 2 years and their annual transfer volumes are less than $10 million (for more, read this SaveOnSend article).
When it comes to a method of sending and receiving remittances, there is also some confusion. Various definitions like “digital”, “online”, “mobile” or “mobile payments” are often use interchangeably. “Digital” is an umbrella word that could include everything or just “online” and “mobile”. Migrants preference for a particular channel is also highly tied to its ethnicity even among neighboring countries:
Overall, digital channels continue being dwarfed by cash transactions. While 70+% of Western Union customers have a bank account, its digital portion of consumer revenues is 9%. The reason for this is simple: consumers are quite content with a cash-based offering:
Starting 2016, incumbents have been attempting to find a middle ground between cash and digital options. They introduced mobile apps that allow cash customers to enter all transfer details online and just bring cash to an agent without filling any additional paperwork. So far, there is no evidence that such efforts generate any material uptake among their customers.
“Mobile money” term typically implies that money are paid from a customer account with his/her telecom provider. Despite being piloted by Western Union in 2007, mobile payments remain a tiny portion of global remittances mostly used for transfers to few African countries like Kenya and Tanzania. It took off in those countries because they didn’t have a high quality bank-card infrastructure for payments like in more developed countries. The most famous examples is M-Pesa launched in 2007 by Vodafone. Remittances volumes among those countries in Africa tend to be relatively small and are, thus, outside of focus for digital expansion by incumbents or Fintech startups. As the result, a mobile money method could be the most cost effective options for these corridors when compared to cash-to-cash sending:
Finally, there is no real difference between “online” and “mobile” since most of online remittances are now sent via mobile devices (60% for MoneyGram in July 2016, 65% for Western Union in December 2016, 75+% in April 2017):
Before discussing differences in operating models across providers, let’s start with a brief context about the key driver of that variability: customer segments targeted by providers. As we reviewed in the article on Western Union, the same provider could have a vastly different approach when targeting various migrant groups. The reason behind it is that each migrant group is unique. For example, consider this graph that compares income and education by ethnicity in USA and try guessing which group is more likely to send money online:
More educated consumers are more likely to send money online for few primary reasons: they tend to be more digitally savvy, they are more open to trying a new provider, and they get paid a taxable income into their bank account. Therefore, 80+% of remittances from USA to India is sent that way versus less than 10% for the USA-Mexico corridor. A reverse relationship is also true: online senders tend to be of higher income. Even within the same migrant group, an average amount sent online could be 50-200% higher than via a cash agent. Higher income also typically results in a different reason for sending money home. Low-income migrants typically transfer smaller amounts ($200-300) monthly to cover basic needs of their families back home or for emergencies. The ones with higher income tend to send more money ($1,000-2,000) but less frequently to their savings accounts back home or as a gift or an emergency.
There are also cultural differences across migrant groups. For example, Indians in USA tend to be one of the most price-sensitive consumers. They are more likely than others to look for the best deal using comparison applications like SaveOnSend and are more likely to switch providers in order to save on fees or to get a higher exchange rate. On the other hand, Mexicans or Filipinos tend to be less price sensitive and are more loyal to their existing provider.
Even within the same migrant group of the same income level, there are behavioral preferences. Some might prefer to pay extra in order for money to arrive immediately, some don’t mind waiting few days in order to get the best exchange rate. All of the above plus many other factors create direct implications for providers on how to market and price their services, which, in turn, drives how they make money.
While startups tend to be secretive about their financial and operational metrics, remittance incumbents, as public companies, have to share those. Luckily for you our readers even startups if they are based in UK and are above certain size must file a basic report, like here for TransferWise:
More detailed reports from Western Union, MoneyGram, Ria Money Transfer (division of Euronet), and Xoom allow us to understand major components of their operating models. Please review below a high-level financial picture of Western Union and Xoom for 2012-2014 – see if you notice major differences and similarities in nomenclature:
Customer Fees and FX Markup
While each provider has its own nomenclature, Western Union and Xoom have a lot in common. A typical money transmitter has 2 sources of revenue: fee and FX markup (or as Western Union calls it, “foreign exchange revenue”). As we discussed in another SaveOnSend article, providers apply radically different tactics within those two revenue components. On the one extreme, Western Union and MoneyGram are trying to maximize both fee and FX revenues not just for each destination, but for each small change in a transfer amount and for each send-receive method. On the other end of the spectrum, TransferWise has a more consumer-friendly approach: charging two-three variations of fees across amounts and, on average, no FX markup:
To make matters even more confusing to customers, most of well-known providers change prices on a daily basis, so a transfer fee could be 2-3x up or down at any point:
Some providers also run a temporarily pricing promotions, hoping to quickly grow a market share in a particular corridor:
As the result, revenue per transaction could be dramatically different for the same provider across destinations and send-receive methods. With prodding from well-funded remittance startups, we frequently see facts-free articles on how expensive incumbents are. Such misleading commentary (see more in this SaveOnSend article) conveniently compares “apples and oranges”: startups’ margins for sending large amounts online vs. what incumbents charge for sending small amounts via cash agents. If we had a bias against incumbents, it would be easy to feel self-righteous by cherry-picking Western Union and MoneyGram’s fees for certain transfers. Just look how much they charge for sending $20 via a cash agent from USA to Mexico: almost 30%!
But if you want to be more objective, you might notice lack of any of the top remittance startups in the above table. While it is easy to talk about helping “poor”, it is actually quite expensive to setup a cash-agent network which will remain, likely for decades, a predominant method of sending money among migrants. In fact, only around 10% of remittance volume is currently sent digitally. That ratio is growing at 15-20% per year, but a larger part of this growth is coming from former wire transfer customers rather than from offline users. Moreover, we know that very few consumers actually send such small amounts and many do it knowing that they could save by a) sending more, b) switching from cash agents to digital sending method.
So how do “old” and “new” providers compare for a typical transfer amount which is sent digitally? Here is a global poster child, ground zero of online consumer remittances, the USA-to-India corridor, unique due its highly educated senders:
Incumbents like Ria Money Transfer and Western Union are among the cheapest providers while some “inexpensive” startups are charging much more. Play with SaveOnSend app to compare incumbents vs. startups margins for other destinations.
Because each provider has a unique mix of corridors it supports and a different mix of online and offline senders (latter is much more profitable), across providers, an average revenue per transaction would be also quite different. While the weighted-average margin for the whole remittance industry is around 6%, Western Union’s global margin is ~5.5%, mostly generated from transaction fees (see full report here):
MoneyGram’s global margin is below 5%, and for Ria Money Transfer, 4th largest global player, it is less than 4%:
For digital-only providers, the gross margins are smaller, but depend heavily on their business model and pricing strategy, for example,
Fixed and Variable Costs
On a cost side, as you remember from financial statements, there are variable costs associated with each transaction and overall fixed costs.
Variable costs include many standard components not specific to a money transfer business:
Let’s cover in details two more unique expenses: paying to receive funds and paying to discharge money. These costs could comprise up to half of ALL expenses of a traditional remittances provider (see Western Union’s 10K for 2014 or MoneyGram’s analyst call about 2015 results)
– Receiving funds: money transmitters are paying agent networks for collecting cash transfers and to banks for sending money from a customer’s bank account or from a linked debit/credit card.
Bank charge relatively little for transfers from their customers accounts, ~30 cents, since it is facilitated through FedACH Service with nominal fees (see “origination per item” in the table below):
Payment for card funding is usually variable, 1-2% of the transfer amount. That is why, consumers who want to use a debit or credit card for funding their transfers are typically charged a much higher fee (credit card funding method has the highest fees in order to cover its higher non-payment risk).
For a bank-funded transfer, fees are usually fixed for big ranges in the amounts (e.g, same fee for sending $0-1,000 or even $0-2,999).
For a cash agent or card-funded transactions, fees would typically change for smaller increases in transfer amounts (see example below or play with SaveOnSend app):
Some money transmitters are trying to minimize the card-based funding cost by using so called “on-us” technique. They are accomplishing it by applying to be a processor of debit-credit card transfers in agent locations. In a sense, they are closing a loop between sending money from a linked debit-credit cards and processing such transfers internally without leveraging Visa/MasterCard/etc. networks.
Payment to cash agents is much more nuanced and covers both fees and FX markup. Negotiated terms could be a split of those or a minimum threshold that a provider must receive while allowing an agent to markup fees and exchange rates at its discretion. Depending on a provider, a retail chain, and a market, there is a vast difference in how revenues are split between these two parties. For example, a market leader like Western Union could be more selective, only paying 10-30% of collected fees to a cash agent. A smaller provider might split fees 50/50, and in some cases, when a money transmitter wants to quickly gain market share, it could add an additional incentive for cash agent networks by promising to pay them up to 100% of collected fees. There is a similar logic when it comes to splitting an FX markup. From the opposite perspective, a larger retail chain could negotiate better terms than a smaller one.
– Discharging funds: providers are paying to banks and cash agent networks to discharge money/give cash to customers. These payments are usually fixed, e.g., a provider would pay $2 per transaction to a bank in a destination country.
For both receiving and discharging process, incumbents and TransFast built direct relationships/connections with large banks and cash networks. On a daily basis, those providers would send funds to a correspondent bank for each destination. For example, if Western Union transfers ten million dollars daily to Mexico on behalf of its customers, it would send that daily amount to its correspondent bank in USA (in dollars), so that bank could then distribute funds in Mexico. Since an exchange rate could move significantly during a day, some providers would also have an in-house hedging-trading desk staffed with few people responsible for buying and selling currencies to minimize an exchange rate exposure. Such daily wires for each destination plus trading staff seem like a significant expenditure, but not in a context of million dollar volumes ($30 wire cost for $1M daily volume with 5% average gross margin represents less than 0.1% of revenues). That is why, Bitcoin-blockchain method of money transfer while theoretically eliminating a need for wire transfers is saving relatively little and actually adding higher additional costs due to extra currency conversions and much higher volatility (for more details on Bitcoin for remittances, read this SaveOnSend article).
Other money transmitters have been deploying its transfer services via intermediaries for some or all markets. In some cases, Xoom uses Earthport, TransferWise – Earthport, WorldRemit – Earthport and BTS, Azimo – CurrencyCloud, Viamericas (aka, Vianex) – Earthport. Such approach is easier-faster, but it might be more expensive in the long run. For example, CurrencyCloud is charging around 0.1% of transfer volume for its services (see full report here):
Using such intermediary also creates a higher risk of dependency on a single provider for the most critical element of back-office processes: transferring funds across entities.
Besides intermediaries like Earthport or Currency Cloud, Fintech startups could also partner with banks or even with direct competitors among incumbents in order to get a faster access to distribution. For example, in December 2016 WorldRemit signed such agreement with Xpress Money (read more here):
“We want to offer our customers the widest and most convenient choice of payout options. Xpress Money is a trusted and dependable money transfer brand with a fantastic network of agents across the world. Our partnership will extend our footprint into new territories and will enable even more people to make secure, instant money transfers.”
Fixed costs include variety of components, many of which are not specific to remittances business. For example, all digital remittance providers have more-or-less standard systems architecture that one could fine in any retail financial company:
Let’s cover the unique components of fixed costs:
Licenses in a state where a money transmitter operates. Getting licensed is a time-consuming and expensive process. It is then followed by a thorough annual audit that looks at all aspects of operations including cyber security. That is why, you see startups like WorldRemit starting with just few, typically smaller, US states where it is easier to get licensed and then working their way up to the largest states over the course of many months if not years. Others, like TransferWise, enter USA by partnering with a licensed provider like PreCash (it later switched to CFSB):
Fraud: 1) “NSF” (Not Sufficient Funds) is when a customer sends money from a linked bank account or card without having those funds. 2) “Misrepresentation” is when a customer lies about NOT sending money after the transfer is completed (according to the US laws, a customer has up to 12 months to dispute a transfer), 3) “Account takeover” is when a customer account is compromised
Here is how the fraud economics works for a $1,000 money transfer from a provider perspective:
average profit per such transaction is ~$5.
so a remittance provider needs 200 successful transfers to make up for one fraudulent transaction (i.e., 0.5% is break-even for revenues)
the overall fraud attempts are 1-5% in total across various migrant groups
money transmitters employ various techniques to keep actual fraud much lower, close to 0.1% (in 2017 FTC’s lawsuit against Western Union, the company’s own complaints database included around 50 thousands fraud-related complaints per year which is around 0.02% of around 250 million transactions per year; but the firm estimated that the actual number is likely five times higher due to under-reporting)
The fight with fraud is constantly evolving as fraudsters are becoming increasingly sophisticated. Some providers have excellent fraud-prevention teams with deep statistics backgrounds and still get ripped off.
Cyber security: similarly to Fraud, preventing hacker attacks is becoming increasingly challenging but imperative due to high potential costs (read about Xoom’s debacle in December 2014).
Compliance: on top of implementing above control-related processes, remittance providers must prove that those processes are working well.
Compliance is VERY expensive. It costs Western Union $200 million per year and involves 2200 employees or more than 20% of workforce dedicated to compliance. Compliance rules could be different in each country which makes them hard to follow, especially, since most of local partners are independent organizations. Imagine the challenge of ensuring that all employees of all your agent partner locations (500,000+ for Western Union) are strictly adhering to compliance procedures.
So despite such massive and on-going spend in this area, Western Union was hit with $568 million fine in the US and ~$2M fine by Ireland’s Central Bank, MoneyGram paid $13M in fines in USA and two back-to-back fines, $0.4M in total, to The Australian Transaction Reports and Analysis Centre. Incumbents often feel that they have been disproportionately singled out for such compliance enforcement. It might be true, but it doesn’t mean that established providers like Transfast or smaller startups don’t take compliance seriously.
All providers tends to use same channels for acquiring customers: Paid Search, SEO, PR, Billboards, Online-TV-Radio-Print Advertisements, Affiliate Marketing, Social, Referrals, Promotions.
Paid Search: a money transfer provider pays to a search engine like Google for an advertisement at the top of the search page. The closer it to the upper left corner the more it costs.
SEO: the goal is to rank as high as possible when customers search online for money transfer options. In the image below, Xoom has both the top SEO result and a paid ad in the right column
PR: goal is to get favorable mentions in the media that look like a legitimate reporting. By definition, an outsider can never know for sure if a particular article was “sponsored” or not. For example, read this article rooting for TransferWise: Richard Branson, Peter Thiel take aim at Western Union. As you could see, it is really hard to find the line between a lazy reporting and a great PR job (best described in Paul Graham’s classic, The Submarine).
There are various ways to arrange PR publications. Sometimes, a publisher would overtly quote a price, e.g., “$60K for 4 positive articles.” Other times, a remittance company would hire a PR firm that would deal with a publisher. In our early review of articles on remittances, the great majority seem to be written under some form of quid-pro-quo or by really lazy-mediocre reporters. If you come across a legitimate quality reporting on consumer remittances, please leave a link in the comments section… Unfortunately for consumers, this trend is likely to expand as brilliantly noted by John Oliver.
Billboards: self-explanatory (see example below)
TV Ads: also self-explanatory (see a typical theme in this ad for WorldRemit)
Affiliate Marketing: the goal is to acquire customers via websites that are popular among frequent senders. When users click on a link displayed on such website, they are transferred to a money transfer provider. To smaller websites, a remittance company would pay a fixed amount (e.g., $2-8 for each new customer) and use an aggregator (e.g., Impact Radius for Ria). Considering that across all channels providers spend up to $50 to acquire a new customer, those payouts seem quite reasonable. Large websites could negotiate customized terms and work directly with providers.
Social: this channel allows remittance companies to leverage social tools (Facebook, Twitter, Google+…) to promote its brand and services. Each provider has its unique approach for balancing between those two components. For example, Western Union tends to focus mostly on branding with pictures of families and food.
Facebook has become an especially potent channel due to its unmatched audience size, ease of targeting, and support for marketing analytics. TransferWise’s uniquely provocative messaging is a better fit for social channels versus competitors as it is designed to drive more conversation/virality:
Fish to the face, electric shock, mustard attack. Watch these people get hit with the unexpected. Luckily, you can send money abroad without the shock of hidden fees: TransferWise.com/phew
Posted by TransferWise on Monday, August 10, 2015
TransferWise targets a 12-month payback across acquisition channels. To accomplish it, TransferWise relies on extensive testing and analytics:
“… we use a load of different sources of data, Lift Tests, Brand Lift, Facebook’s pixel data, our own data and attribution as well as some 3rd party data.”
Referrals: the goal is to motivate existing customers to act as sales agents for a provider.
Here is a visual example of such referral programs’ effectiveness, courtesy of TransferWise. According to TransferWise’s description, it was done for 3 countries, in different colors, with each dot representing a TransferWise user. Each connected dot is a user they invited (or were invited by). TransferWise claims that around half of new customers come through referrals.
Promotions: to attract new customers, remittance providers would offer a free 1st transfer or a payment/gift card for the first transaction.
A more recent promotional tactic, started in 2016 by Transfast and Remitly offers a very favorable exchange rate for the first transaction:
Valuations across Incumbents and Startups
At SaveOnSend, we don’t have any expertise in valuing a company, so we will only present facts and would look for an expert opinion to explain a potential misalignment.
Let’s compare all providers based on the latest revenues and valuation data:
It seems that revenues of startups are valued 5-10x higher than revenues of Western Union. Why is that? It is not because startups generate much better margins. Since they don’t have a high-margin cash business and are in the aggressive build-up phase, startups are much less profitable in a relative comparison with incumbents. Such disparity in valuations is obviously due to underlying growth expectations among investors. Read explanations from lead investors in the previous funding round for: TransferWise – Learning From My Mistakes, Remitly – On a Mission to Improve, WorldRemit – Drive Global Growth.
Here is the core of their common investment thesis:
“significant disruption” “clear shift to online-mobile”
“no innovation” from incumbents
startups offer much more “convenient, low-cost solution”
Of course, Fintech startups enabling this line of thinking with their predictions:
The reality of 2017 is “NO”, “WRONG”, “NO” to above investment thesis. The shift to online is crawling at 1-2%, revenues of top providers continue growing, incumbents have same online-mobile tools as startups if not better (read our post on Western Union), startups offer less convenience with limited destinations, send-receive methods, and, sometimes, longer delivery of funds, startups are often more expensive than incumbents (remember that “3.8%” mentioned above for Ria Money Transfer includes a much larger portion of offline transfers), their compliance costs might be lower in absolute terms, but it is mostly due to a small size and lack of cash agent network.
Let’s look at the growth in digital transfers:
You might be surprised to discover that incumbents are ahead of most of Fintech players. However, startups keep growing at a faster rate for now:
But how such super successful investors could be so out-of-touch with reality? Besides arrogance and laziness, there is a more prosaic reason: this narrative makes a good intuitive sense (read our TransferWise post for more on that). With the rarest exceptions, every “expert” and publication repeats this narrative when writing about remittances, so it is naturally hard to be open-minded that “everybody might be wrong.” At SaveOnSend, we also believed in those points prior to launch and it took us few months of continuous hands-on research to discover a counter-intuitive and complex reality: like the fact that compliance and anti-cyberfraud costs have skyrocketed in the last few years, or that incumbents are often less expensive than startups, or that Western Union went online in 2000, began piloting mobile payments in 2007, and had a mobile app in 2011 (that would also imply that Western Union should be considered a pioneering FinTech player that probably did more for connecting “underbanked” than all remittance startups combined).
So the reason for much higher relative valuations might be due to these startups’ investors being simply as out-of-touch with realities of consumer remittances as the rest of us.
To be clear, at SaveOnSend we think it would be highly beneficial for consumers if these startups survive and become stable profitable businesses. Even when charging higher prices, startups create competition, thus, motivating incumbents to be more proactive. It is also awesome that funding from wealthy VC owners creates jobs and is partially redistributed to some low-income recipients across the world. Our specific issues with these remittance startups are inaccurate statements about their offering vs. incumbents and hypocrisy of their claims to “help poor” while targeting better-off online senders in the world’s wealthiest countries.
Back to valuations… so what would be a fair value of these startups? They have similar tools and use similar acquisition channels as online divisions of incumbents. Moreover, startups are relying on 3rd parties for key processes and, thus, don’t necessarily have lower costs and risks. So it really comes down to a) management, b) funding. If startups have better management than incumbents, they could evolve faster. If FinTech bubbles stop happening, startups could keep raising money and keep acquiring customers at a loss.
Finally, let’s assume that by focusing too much on facts we are missing a subliminal long-term trend and there is indeed a massive upcoming shift to digital remittances. So what would then the industry look like using the world’s most advanced online corridor, USA-to-India, as a preview of things to come?
$600B industry at 1% gross margin and 0.2% net margin = $6B in revenues and $1.2B in profits
How much would you pay for an annual stream of $1.2B which could further decline? $10B?
In that case, using today’s valuations, it would imply ~10% global market share for TransferWise in a foreseeable future, ~5% for WorldRemit, and ~1.5% for Remitly.