Engagement & Retention project | Lending Stream

📄

Engagement & Retention project | Lending Stream

Lending Stream - An Overview

Lending Stream is an online short term loan product designed specifically for Brits with less than perfect credit score (low credit worthiness). It is mainly a 6 month loan with some users eligible for a 12 month loan term too.

A user needs to apply for it on our website by filling up a quick online form. Once its submitted, we run our underwriting and affordability assessment to decide whether or not to lend money, which takes about 5 seconds on average. If the application is approved, the user needs to add a bit more information and review the loan agreement after which the money is transferred to their bank account within 90 seconds.

Lending Stream's value proposition, at a very high level, is their ability to offer short term loans to users who otherwise would find it difficult to borrow from banks and other financial institutions, while keeping them away from getting into debt-trap and doing all of this in a couple of minutes. Lending Stream is regulated and authorised by Financial Conduct Authority (FCA).

Let's understand Lending Stream as a product in more detail💡

Problem that it solves:

Its a short term loan product designed mainly to lend small sum of money to people with bad credit history, who find it hard to get it elsewhere quickly. Some good/fair credit people also borrow from us when they need the cash urgently and have nowhere else to go.

People refer to such loans as Payday Loans (an older term which isn't right in the current context), Short term loans and bad credit loans. These products are regulated by the Financial Conducts Authority (FCA) under High Cost Short Term Credit (HCSTC) category. These customers are majorly categorised by their credit worthiness and their bankability - near sub-prime and non-prime.

Basic features:

Fully online product where-in you apply online through our website that takes a few minutes.
Instant credit decision (a few seconds) and Superfast payout (within 90 seconds).
Clear loan terms - daily interest rates and late fees.
Easy EMI repayments over 6/12 months.
Option to borrow multiple times (2 loans max at the same time).
Very expensive form of borrowing.
Available only for UK residents who are 18+, employed and have bank account & a debit card.

Product Stage

PMF achieved many years ago. Undergoing mature scaling now. Leader in the market (>30% market share of HCSTC in the UK).

Now that we understand the product really well, let's design our Engagement Strategy:

STEP 1 → Product

1. What is the core value prop of your product? How do users currently experience that core value prop repeatedly?

Our Core Value Proposition is to "Solve urgent cash needs for UK residents by lending them smaller sum of money in minutes, especially when their credit history isn't perfect".
Users experience this core value proposition by getting the cash that they need after getting onboarded as a new customer and then use their good repayment behaviour to keep borrowing from us whenever there is a need (as long as we feel they can afford these loans).
1. They can take the repeat loans by logging into their LS account and applying from there. They'd be able to apply for another loan only if they have shown positive repayment behaviour in their last loan(s). They can hold 2 loans parallely at a time with us.
2. They can make the repayments using Auto Debit feature OR manually by loggin into their account.

2. What is the natural frequency of your product?

Casual User: 1-2 loans in the a 2 year window
Core User: 3-5 loans in the a 2 year window
Power User: 5+ loans in the a 2 year window

3. Do you have other sub products? What is their natural frequency like?
No, there is only one product that is the loan product, however, a key feature of a loan product (also a core action) is making repayments which in almost all cases is paid once a month as long as there is an active loan.

4. What is the best engagement framework for your product?

Engagement Framework	Relevance	Key Metric to be tracked	Rationale
Frequency	Primary	X loans over Y months	We'd want our users to borrow from us as many times as they need. This will help them realise the CVP of our product by paying off an emergency bill whenever that occurs regardless of amount. More loans == more no. of times that customers realised CVP == Higher brand consideration == Longer term retention == High LTV. You must now be thinking, why not optimize for the actual amount they borrow as higher loan amount == higher profits ..., You're right but there's a catch. Read the next section to understand.
Depth	Secondary	£X over Y months	Our customers don't control the amount that they're approved for and can only control how much out of the approved amount, do they borrow. In most cases, our customers take out the entire approved amount leaving lesser scope to increase the average loan size. Having said that, we should make UX optimizations to increase the {approved amount} to {funded amount} ratio to increase the LTV.
Breadth	Nill	-	We offer only 1 product i.e. a loan. There are no other products/sub-products/features that would help the customers realise more value proposition.

Step 2 → Define who is an active user of our product

For our loan product, users get activated when they borrow (and successfully close) X loans in the first Y months. However, a user once activated can become dormant and even churn.

So, a good way to define active users would be to look at:

how recently did they borrow
how frequently did they borrow

Now, customers might not be able to borrow until they're deemed affordable by our underwriting but they could still be active on our the product by making repayments for the open loans, therefore it would be less appropriate to look at "time since their last loan" to call them active.

Instead, the appropriate definition for an active user could be:
"A user having a non-zero outstanding balance over the last 12 months and haven't gone bad (default, chargeback, etc.)"

We decided on 12 months in the above statement based on the natural frequency of our users.

Step 3 → Customer Segmentation

Before, we get into our segmentation, let's first understand our users. And the most efficient way to do that is by tabulating the traits of our core ICPs.

Ideal Customer Profile (ICP):

Name	Chloe	David	Sam	Matthew
Persona	Bad Credit borrowers - Young employed	Bad Credit borrowers - Older employed	Bad Credit borrowers - Self employed	Emergency borrowers
% share of current customer base	60%	15%	5%	10%
Age	18-44	45-65	18+	18+
Monthly Income	£1000 - £3000	£1000 - £2500	£800 - £3000	£1500 - £4000
Occupation	Full time/Part time employed A good share - NHS workers, and superstore employees (Tesco, Sainsbury, etc.)	Full time/Part time employed	Self employed A good share - gig workers like Deliveroo and Uber drivers	Full time/Part time employed A good share - NHS workers, and superstore employees (Tesco, Sainsbury, etc.)
Need	Quick - short term loan	Quick - short term loan	Quick - short term loan	Urgent cash
Pain Point	bad credit history - limited options to borrow	bad credit history - limited options to borrow	bad credit history - limited options to borrow	Other loan options take time
Brand loyalty	Low - go with cheaper option	Medium - go with cheaper option	Medium - go with cheaper option	Low - go with cheaper and quicker option
Early adopters	High	Medium	Medium	Medium
Spend time on	Instagram, Facebook, TikTok, YouTube, OTT, Sports TV channels, Radio, Spotify	Cable TV, Radio, Facebook, YouTube	Instagram, Facebook, TikTok, YouTube, OTT, Sports TV channels, Radio, Spotify	Instagram, Facebook, TikTok, YouTube, OTT, Cable TV, Radio
Perceived Value of Brand	Low	Medium	Low	Medium
Frequency of use case	High	Medium	Medium	Medium
Average loan size	Medium	High	Medium	Medium
What do they spend most on	Rent, commute, food and groceries, energy, utilities	Rent, commute, food and groceries, energy, utilities	Rent, commute, food and groceries, energy, utilities	Rent, commute, food and groceries, energy, utilities
Most valued/liked features of the product	Simple application, online, fast decision, fast cash disbursal, customer support, muliple loans	Simple application, online, fast decision, fast cash disbursal, customer support, muliple loans	Simple application, online, fast decision, fast cash disbursal, customer support, muliple loans	Simple application, online, fast decision, fast cash disbursal, customer support, muliple loans
Most hated things about the product	High interest rate, late fees, application rejections, customer support	High interest rate, late fees, application rejections, customer support	High interest rate, late fees, application rejections, customer support	High interest rate, late fees, application rejections, customer support

In the ICP table above, I have highlighted the key traits of our customers that would likely impact their engagement with our product.

After speaking to these customers, looking at their product reviews and analyzing their performance data, I've concluded that most of our customers have very similar reasons to engage with us and the key difference is their frequency of usage.

Time since last zero OPB: This is an important metric as we've used this to define Active users as well in a way. So, when a user has had X days since the last time they had no outstanding principal balance (all loans closed) - higher the X, lower their engagement and lower brand recall. Therefore, while designing the engagement campaigns, we need to treat customers differently based on this metric to have a better impact. E.g. Users with 1 active loan would have a higher engagement as they're still making repayments and are already receiving other service emails like payment reminders, successful repayments, etc. There's no need to target them too frequently.

* Users can have at max 2 loans at the same time, so users with 2 active loans wouldn't be able to borrow until they close at least 1 of them.
* Users with >12 months since the last zero OPB are deemed in-active users - churned or about to churn.

Therefore, segmenting our customers based on their usage frequency (Power/core/casual) X Time since last zero OPB would help create a robust engagement strategy. For ease, I've combined Core and Power into one segment as the same engagement campaigns could be utilised for both of them. All other traits of these users would be same as our ICPs.

Usage Frequency	Casual	Core/Power
Days since last 0 OPB	0 (1 active loan)	0-12 Months	0 (1 active loan)	0-6 Months	6-12 Months
Natural Frequency	1-2 loans in 2 years	1-2 loans in 2 years	3+ loans in 2 years	3+ loans in 2 years	3+ loans in 2 years
Level of Engagement	Temporarily high, otherwise low	Low	Temporarily high, otherwise medium-high	Medium-high	Medium
Recency of use case	High	Low	High	Medium	Low

Step 4 → Engagement Campaign

Now that we have segmented our customers to optimize the engagement strategy, let's design some engagement campaigns.

I've designed 5 campaigns targeting the customers as segmented in the earlier section. You can find the campaign flow here.

Please note that we run an offline eligibility on a daily basis to list all the customers who are very likely to be approved if applied today (based on their last application details). These campaigns would only cater to customers who're eligible for a loan on the day of drop.

One thing to note here is that the messagig could be customised within each campaigns for different set of users:

Own marketing driven users Vs Partner channel driven users -- More brand focus for Partners-led users to increase TOMA
ICP based content to improve the communication response -- More casual tov for Chloe vs more easy to understand and direct messaging for David.

Engagement Campaign	Goal of the campaign	Persona/Type of User	Rationale/Additional info	Offer	Pitch and Content	Channel of Distribution	Frequency and Timing	Metrics to track	Success Metrics
Campaign 1	To inform users that they can have 2 loans at once and to trigger an action (take another loan). Convert them into Core users.	Casual users with an active loan	• A casual user might not be aware of topup loans, so sending this would increase loans per customer. • A high freq of targeting could upset them as they have lesser use case for borrowing and would also be engaged via other service emails and them having to make repayments- so kept it once a week. • Test different day/time combination to find the best one	No special offer except that they are eligible to borrow more.	Subject: Did you know that you could have another loan with us? Hey there, do you know that you could have 2 loans with us at the same time? We understand that sometimes, a topup on your loan is what you need. CTA - Apply now/Topup your loan	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • % Users moving into higher engagement cohort/more usage • Unsubscribe rate	• X% increase in loans per customer • Y% casual users become core users
Campaign 2	To disburse more loans to these existing customers. Convert them into Core users.	Casual users with {0-12 months} since last zero OPB	• A casual user might not remember the brand, so sending this would increase loans per customer. • A high freq of targeting could upset them as they have lesser use case for borrowing - so kept it twice a week. • Test different day/time combination to find the best one	No special offer except that they are eligible to borrow more.	Subject: Lending Stream's here when you need another loan Hey there, we were happy to help you with your last loan and we’re here if you need another loan. You’re eligible to borrow up to £1500 now. All you need to do is login and apply. CTA: Apply now	Email + SMS	2 times a week (Tuesday, Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • % Users moving into higher engagement cohort/more usage • Unsubscribe rate	• X% increase in loans per customer • Y% casual users become core users
Campaign 3	To disburse more loans to these existing customers. Convert/retain them into Power users.	Core/Power users with an active loan	• A core/power user with an active loan might need another loan, so sending this would increase loans per customer. • A high freq of targeting could upset them as they would be engaged via other service emails and them having to make repayments- so kept it twice a week. • Test different day/time combination to find the best one	No special offer except that they are eligible to borrow more.	Subject: Need another loan? We're here to help Hey there, glad to see that you’re finding us helpful. As you know you can have 2 active loans with us at a time. So, if you’re feeling the heat, you are eligible to topup your account with another loan with us. CTA: Apply now/Topup your loan	Email + SMS	2 times a week (Tuesday, Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • % Users moving into higher engagement cohort/more usage • Unsubscribe rate	• X% increase in loans per customer • Y% core users become power users
Campaign 4	To disburse more loans to these existing customers. Convert/retain them into Power users.	Core/Power users with {0-6 months} since last zero OPB	• Core/power users tend to borrow whenever they need and informing them when they're eligible would increase loans per customer. • A higher freq of targeting should be fine as they have frequent use case for borrowing - so kept it thrice a week. • Test different day/time combination to find the best one	No special offer except that they are eligible to borrow more.	Subject: Congrats, you're eligible to borrow up to £1500 Hey there, glad to see that you’re finding us helpful. As always, we’re here if you need another loan. CTA: Apply now	Email + SMS	3 times a week (Tuesday, Thursday, Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • % Users moving into higher engagement cohort/more usage • Unsubscribe rate	• X% increase in loans per customer • Y% core users become power users
Campaign 5	To disburse more loans to these existing customers. Convert/retain them into Power users.	Core/Power users with {6-12 months} since last zero OPB	• Core/power users tend to borrow whenever they need and informing them when they're eligible would increase loans per customer. • A higher freq of targeting should be fine as they have frequent use case for borrowing - so kept it thrice a week. • Test different day/time combination to find the best one	No special offer except that they are eligible to borrow more.	Subject: Good news, you're eligible for another loan! Long time no see? That’s a good news, this means you’re able to manage your finances well. As always, we’re here if you need another loan. CTA: Apply now	Email + SMS	3 times a week (Tuesday, Thursday, Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • % Users moving into higher engagement cohort/more usage • Unsubscribe rate	• X% increase in loans per customer • Y% core users become power users

Step 5 → Design some Product Hooks

There are a few product hooks that we can design that'd increase the user engagement on our peoduct and also help with resurrection of currently churned users. The key goal would be to convert all new casual users into core and core into power.

I've designed 3 product hooks below in brief. More details to follow at a later time.

Loyalty program - Design your interest rate in a way that a better user repayment behaviour and their engagement with the product adds more value to them. E.g. Interest rate that decrease after every 3 loans.
Reward good repayment behaviour - Design a program to reward users on their good repayment behaviour like closing a loan on/before time. Thus increasing their trust on the brand and hence the engagement. E.g. 10% cashback of your entire interest paid if you close the loan on/before time. Please ensure this doesn't end with us a net LTV loss as closing the loan too soon would mean lower revenue.
Add new features/subproduct - Add a subproduct or feature that would add more value to the user only they are your active users (you can decide to open it for all too). E.g. Partner with a discount coupon merchant to let them use the discount coupons that are not available to public. For this, they need to use some coins that they earn when they borrow/make timely repayments, etc.

And that'a wrap for our Engagement strategy.

Retention Design

Step 1 → Understand our retention

Since it is a longer term and a lower frequency product, we would need to understand the retention for our product by looking at how many customers are still active after say every 6 months.

As defined earlier, A user having a non-zero outstanding balance over the last 12 months and haven't gone bad (default, chargeback, etc.) is an active user of our product.

So, if a good-standing user had at least one day when the outstanding principal balance was >0 during the last 12 months after the X months of acquisition, the user would be termed as a retained user for MX.

* Good-standing user - A user who has not defaulted on the loan or showed any signs of other bad behaviour. If a users misses payment but gets back on the track before the loan schedule, we could consider him/her good-standing.

Bird's-eye view

1. What is your current retention rate in terms of users?

1. MAU (monthly active users) - ~2% >> we see about 20K unique users from our 1M user base who borrow in a month.
2. YAU (yearly active users) - ~8% >> We see about 80K unique users out of our 1M user base who borrow in a year.

These metrics are low as we've had this product for over 15 years and a lot of our older customers are no-longer our target customers plus there have been instances of poor product experiences in the past leading to even lower retention from those customers. Another reason is regulatory requirements to not lend customers more than a certain times over a stipulated time and also over a lifetime.

If we were to look at these metrics for recent cohorts, the numbers would be much better.

Another interesting thing that you'd see in the above metrics is that the YAU is just 4 times of MAU - which means, there are sets of users who borrow multiple times in a year and sets of users who don't borrow at all (churned).

2. At what time period does your retention curve flatten?

Retention curve (i.e. had >0 OPB over the last 12 years of X month from acquisition and are good-standing)
* Guesstimate based on looking at similar data for the last few years - never really looked at this data.

1. 1. M1- 100%
  2. M12 - 93%
  3. M18 - 80%
  4. M24 - 65%
  5. M36 - 35%
  6. M48 - 25%
  7. M60 - 20%
  8. M72 - 15%
  9. M84 - 10%
  10. M96 - 5%
  11. M108 - 4% and so on...

So, based on the above data - the retention curve kind of flattens at about 8-9 years since acquisition. Post that only super loyal users stick.

Microscopic View

1. Which ICPs drive the best retention?

Both Chloe (higher use case but Lower loyalty) and David (lower use case but higher loyalty) show similar retention rates. Other minor ICPs also follow a similar pattern. Overall all the 4 ICPs have casual/core/power users who have varied retention rates.

2. Which acquisition channels drive the best retention?

Partner channels intuitively drive lower retention than non-partner (own marketing) channels in the start as our onw-marketing efforts focus a lot of presenting ourself as a superior product in the market while the partner leads are essentially leads who instead of applying directly with us, applied with lead brokers displaying lower brand loyalty.
However, this gap reduces over time and more importantly the average loan size for partners-driven users is higher than own-marketing driven users ultimately making both the channels similar in the longer term.

Step 2 → Reasons of Churn

After speaking to a lot of customers who have churned and are showing signs of churn plus reading through numerous negative reviews on the review sites - I've put together the key reasons of churn and their proposed solutions in the following mind map. Access the whimsical here. There are also a couple of involuntary churn reasons (not controlled by users) that we as a business could do something about.

The top solutions are:

Design a loyalty program that rewards a lower interest rates to users based on their history with us. -- Product hook
Offer a one-time cashback reward on repaying the next loan on time. -- Product hook
Add a new financial wellness feature/product to the website and provide access to existing users. -- Product hook
Send emails to users dropping from the open banking journey making them realise why is this a necessary step and improve their trust in the brand and the process by adding testimonials, regulatory requirement and such.
Ad-hoc email campaigns informing sets of users who had poor experience with the product/service lately with an update on their pain point plus a one-time cashback reward on repaying the next loan on time.
Introduce different products to users who are not our target customers anymore and earn a commission.

Negative actions to look out for:

Step 3 → Resurrection Campaigns
Use this link for a better readability.

Resurrection Campaigns	Goal of the campaign	Persona/Type of User	Rationale / Additional info	Offer	Pitch and Content	Channel of Distribution	Frequency and Timing	Metrics to track	Success Metrics
Campaign 1	To resurrect the users and let them borrow again.	Casual users who have churned (0 OPB for >12 months)	Users likely churned as they found a cheaper alternative. Offering them a lower interest rate on their future loans could get them back. If they are churned because they no more need the loan. You'd still want to keep them engaged, so that they come to you for their next loan.	Loyalty program. Lower interest rates.	Subject: Interest rate SLASHED. Apply now HI, we've not seen you for a while. You're eligible for a loan and to make the deal sweeter, we're introducing our loyalty program exclusively for customers like you. Take out your next loan at X% lower interest rate. Here's how it works... CTA - Get your loan now	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • Unsubscribe rate	• X% customers targeted take a loan
Campaign 2	To resurrect the users and let them borrow again.	Core/Power users who have churned (0 OPB for >12 months)	Users likely churned as they found a cheaper alternative. Offering them a lower interest rate on their future loans could get them back. If they are churned because they no more need the loan. You'd still want to keep them engaged, so that they come to you for their next loan.	Loyalty program. Lower interest rates.	Subject: You're gonna love our SLASHED interest rates HI, we've not seen you for a while. You're eligible for a loan and to make the deal sweeter, we're introducing our loyalty program exclusively for customers like you. Take out your next loan at X% lower interest rate. Here's how it works... CTA - Get your loan now	Email + SMS	Twice a week (Tuesday, Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • Unsubscribe rate	• X% customers targeted take a loan
Campaign 3	To resurrect the users and let them borrow again.	Users who didn't continue with the new open-banking journey when it triggered.	These customers likely didn't come back to apply as they didn't want to go through the open banking journey that needs them to login to their net-banking while applying. If you make them feel that this is a new yet necessary step and get them to trust you, they could come back. These users would have gotten emails after they dropped from the journey informing why is Open banking needed, but a monetary reward alongside a bit more detailed explanation could do the job.	10% of their interest amount as cashback if they close their next loan on/before time.	Subject: 10% off on your next loan with us Why is Open Banking necessary? Hi, we noticed that you didn't complete your last application process with us. We understand that Open banking was a new step that you might not be comfortable with - but here's why everyone in the industry is doing it (link to a blog with an infographic attached in the email). As you're a valuable customer, we'd love to give you a cashback of 10% of your total interest paid if you close the loan on/before time. CTA - Apply now	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • Unsubscribe rate	• X% customers targeted take a loan
Campaign 4	To keep some users from churning and let them borrow again.	All users who have raised multiple support tickets/interacted with CS multiple times over the last few days.	These are the users who could churn soon if not tended to. Offer them a discount alongside a assurance that their issue is resolved/they'll not face such problem in the future. It could keep them from churning. Ad-hoc campaipgn based on the reason of their complaints category.	10% of their interest amount as cashback if they close their next loan on/before time.	Subject: 10% off on your next loan with us Hi, we know that you might not have had a pleasant experience with us lately due to ..., we have now fixed this and can promise that you won't face such things again. As a token of our apologies, we'd love you a cashback of 10% of your total interest paid if you close the loan on/before time. CTA - Apply Now	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • Unsubscribe rate	• X% customers targeted take a loan
Campaign 5	To resurrect churned users who are better suited for our cheaper product - Drafty	Users who've churned as they are qualified for better deals outside based on their credit score/history and are eligible for our other product Drafty.	They've churned as they are finding better deals outside due to their improved credit score. Offering them Drafty could make a loyal Drafty customer.	Introduce Drafty (much cheaper line of credit)	Subject: Congratulations! You've got an Upgrade. Hi, As a valued LS customer, we'd like to give you an upgrade. Our sister brand Drafty can provide you a revolving LOC up to XX at a much cheaper rate - Y % APR CTA - get Drafty	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Signups • Unsubscribe rate	• X% customers targeted becoming Drafty customers
Campaign 6	To monetise churned users who are not our target customers anymore.	Users who've churned as they are qualified for better deals outside based on their credit score/history but are NOT eligible for our other product Drafty.	They've churned as they are finding better deals outside due to their improved credit score. An aggregator like Clearscore could find them cheaper credit whenever they need - so introduce them to Clearscore to earn a commission.	Introduce Clearscore	Subject: Congratulations! We've a good news. Hi, we're so happy to see that your efforts have helped you become more financially resilient. We understand that LS is no more a suitable lending option for you, so we have partnered with Clearscore to find you the cheapest credit whenever you need. CTA - Sign up for Clearscore	Email + SMS	Once a week (Saturday) 10-11 AM	•Open Rate • CTR % • Signups • Commission amount • Unsubscribe rate	• X% customers targeted sign up for Clearscore
Campaign 7	To monetise churned users who are churned as we didn't approve them and are likely to not be approved in the future.	Users who are churned as we didn't approve them and are likely to not be approved in the future.	They've churned as we are not approving them for a loan. Ask for their consent to find them a loan with partner lenders.	Let us find you a loan	Subject: Let us find you a loan Hi, we're so sorry that we're not able to lend you due to various reasons. However, we can try to find a lender who can. CTA - Apply now	Email + SMS	Once a week (Saturday) 10-11 AM	• Open Rate • CTR % • Conversion rate • # Loans disbursed • Unsubscribe rate	• X% customers targeted take a loan