
Case 1.1 Human engagement can prevent churn among SMEs
Once usage has dropped, re-engaging a customer can be expensive and often fails. Instead of waiting until a user has churned, it is helpful to identify those who are likely to do so, and motivate them to re-engage before the fact. Doing so requires investing in data analysis and abilities to predict who might churn, and developing effective (re)engagement strategies.

Tienda Pago serves SME retailers specialized in fast-moving consumer goods (FMCGs) in Peru and Mexico via short-term rotating credit for inventory, which allows them to place larger, more diverse orders at more affordable rates. Moreover, Tienda Pago partners with distributors to pay for inventory directly, thereby saving distributors cash management efforts and facilitating their access to smaller merchants. The company has found that if retailers stop using Tienda Pago’s credit product for several months, they are unlikely to resurrect down the road. However, they have found that reaching out to merchants can prevent them from disengaging and bring down churn rates.
Given these observations, the Tienda Pago team developed a data analysis and machine learning algorithm to target stores that are likely to churn, and invests in human-touch engagement to drive usage and create stickiness, thereby recovering users before they churn.
Background
The informal economy in Mexico is sizable, having contributed 26 percent of GDP on average between 2003 and 2012, and employing approximately six in ten Mexican workers. Most of these businesses are small, in fact nine of every 10 businesses in Mexico are SMEs.
Together, the large numbers of small enterprises and high levels of informality, mean that many owners do not have access to formal financial services. Research has found that more than half of SMEs lack a bank account and more than three quarters lack access to credit. Even those SMEs that do have access to credit via a commercial bank only do so via a guarantee scheme from NAFIN, Mexico’s main development bank.
Since access to appropriate formal financial services doesn’t exist for this group of SMEs, most depend on credit from suppliers. This poor access to credit severely hamstrings SME growth and operations since most depend on inventory credit to stock their stores.
While tenures and repayment schedules between SMEs and distributors match (retailers typically repay as they restock on a revolving basis), the effective interest rates are high and such credit locks retailers into disadvantageous relationships with specific suppliers.
Those who do not take credit from suppliers have few alternatives; they rely on informal lenders who evaluate the SME owner as an individual, rather than as a business owner, again at very-high effective interest rates.
About Tienda Pago
Tienda Pago saw an opportunity to serve these SMEs by offering inventory credit facilitated through partnerships with large distributors. The SME retailers use Tienda Pago’s app to make an order for fast-moving consumer goods, for which Tienda Pago extends the store a loan while paying the distributor directly. The retailer then repays Tienda Pago for the order (plus interest) on a weekly basis.
Tienda Pago benefits both retailers and distributors. With access to the Tienda Pago credit, retailers can increase their order sizes, improve the quantity and diversity of their inventory, and potentially generate 20-30 percent more in revenue. Distributors save on cash management since the delivery route effectively becomes cashless. Their delivery teams no longer need to collect payments, which means saved time, fewer errors, and lower risk. In fact, some distributors are now including areas for delivery that they had previously deemed unsafe since cash transactions are no longer unnecessary.
As of July 2020, Tienda Pago counts 45,000 registered users, divided among power users (use on a weekly basis), regular users (monthly), and occasional users (quarterly or less frequent). The challenge for Tienda Pago has been that regular and occasional users churn 30 percent more than power users.
Human touch for regular usage
To address these low usage rates, the team tested a number of engagement strategies. For example, the team tested digital discounts, free transactions, and reminder campaigns (via WhatsApp and the Tienda Pago app). They found that promotions from distributors or Tienda Pago, like a free case of beer, reliably got merchants to place orders but that such promotions were expensive to offer on a recurring basis. Moreover, taken as a group, they found that these strategies achieved an additional order, but none reliably moved the needle to establish more frequent usage. Instead, the team found that many retailers would just take advantage of the promotion and then wait for the next one, but never moved to become regular or power users.
Ultimately the Tienda Pago team realized that meeting or speaking with the owners (either in person or via the phone) was the most effective way to encourage more frequent usage. Regardless of the script used or the staff person involved, Tienda Pago found that users had a similarly positive response to human outreach. However, given the cost of this strategy, the team leveraged a machine-learning approach to target these conversations among users most likely to move into the regular usage category. This data-driven approach has targeted their efforts on two moments in the user journey: 1) establishing regular usage with new merchants, and 2) preventing churn among established users. Human touch, together with this data-driven targeting, has effectively established regular usage and prevented churn among 25 percent of users.

The Tienda Pago team tested a number of strategies to keep users engaged. Rosa (L) and her daughter Emily (R) rely on Tienda Pago to pay for their shop’s most expensive weekly deliveries, which gives them “peace of mind,” says Emily.
Become part of the established routine
A look at the data revealed that power, regular, and occasional users diverge early in their engagement with Tienda Pago; in fact, during the first 90 days of usage. If retailers made weekly or biweekly purchases in the first several months, they would likely go on to become power users of Tienda Pago. More precisely, data analysis showed reaching six transactions in the first 90 days correlated to a high likelihood of that merchant becoming a power user. Such users exhibit this behavior early in their engagement with the service, and very few get there by ramping up from a lower-frequency routine. Furthermore, once a user becomes regular, maintaining engagement is straightforward via WhatsApp and other low-touch mechanisms. This means that Tienda Pago needs to become part of the merchant’s ongoing purchasing routine soon after acquisition.
In 2017, the team found that 57 percent of new customers were reaching the weekly/biweekly purchasing routine (six transactions in 90 days), but that around 40 percent were falling into the low-frequency user groups fairly soon after registering.
Now, to address those who are slow to take up the product, the team carefully tracks usage during the first 90 days and immediately deploys various touchpoints to encourage more frequent usage. The team first tries to contact the merchant via phone, and if they are unable to connect with them, pays a visit in person. They target users who have made at least four orders (so have had some positive interactions with the service) but have yet to become power users.
This targeting has improved uptake. Whereas 57 percent of new merchants became power users in 2017, by 2020 over 67 percent do so, thanks to Tienda Pago’s additional engagement strategies.
Prevent churn
The team also deploys human touch when a merchant may be experiencing a change in their purchasing routine (perhaps due to a change in family circumstances, opportunities, or finances). Reaching out to these users before they drop off helps Tienda Pago prevent churn.
For the second group, those moving from high- to low-frequency usage, the triggering event is usually (around 50 percent of cases) a default. As such, the team works with high-frequency users at their first default to create a payment plan or find some other solution.
The team has also developed a machine learning algorithm that analyzes transaction size, frequency, recency, and other metrics to identify those retailers who are likely to churn based on historical data. Once identified, the team reaches out to remind the merchant about the service and to build the relationship with Tienda Pago.
In deploying these human touchpoints, the team must carefully balance cost and effectiveness. While a number of visits can reliably get merchants back to using Tienda Pago, at some point the cost of these visits can exceed the business upside of the individual account. The team makes an effort to re-engage a merchant for three months. After that period, they deactivate the client, so that, if he wants to reactivate, he has to do a quick training on how to use the app, ensuring the team has a chance to touch base and re-engage that merchant.
As SMEs start reactivating after COVID-related lockdowns, Tienda Pago is readying their operations to help them get back up and running. SMEs will need access to credit to restock their shops and intelligence to anticipate changing needs. In both cases, Tienda Pago will be an ideal partner to ensure their longevity.