Rarely does a business come along that adds value, is inoffensive to all stakeholders, and has a reasonable valuation. Cardlytics provides that and more, making it my largest position both in the portfolio I share on this blog as well as my other investment accounts and those of my family.
Cardlytics is a purchase intelligence platform (“PI”) that connects marketers and bank customers using the bank’s (“FIs”) purchasing data.
Basically, Cardlytics analyzes scrubbed customer purchasing data behind the FI’s firewall and uses that information to show that bank customer highly-targeted advertisements. These are not traditional ads. They are “offers” that come in the form of credit card rewards and discounts. A typical “ad” found on an FI’s mobile app or website looks like what is found at the bottom of the following image, something you may have seen before on your banking app:
Value for All Stakeholders
There are three stakeholders CDLX serves: FIs, bank customers, and marketers. CDLX makes all three of these stakeholders better off without the common pitfalls advertising platforms suffer from.
The banks that integrate the CDLX platform earn a share of the revenue CDLX generates.
First, this is an non-interest revenue stream that comes at no cost to FI’s: it is a productive use of empty real estate on their mobile apps and websites.
Second, the CDLX platform drives credit card spend from FI customers on the macro-level. The interactive, tap-to-earn feature leads to specific credit card usage over non-FI payment means. For example, if the PI platform on Chase offered 5% off my Starbucks purchases this month, I would use my Chase card more often in order to enjoy the offer. This means the mix of swipes between my credit cards would shift toward Chase over Bank of America during the offer period. Additionally, a number of purchases that would have been done in cash shift to the credit card. The Purchase Intelligence platform can also drive higher gross receipts, especially in larger spend categories such as travel, furniture, etc.
Third, the PI platform is a general feature bank customers enjoy and is part of improving their user experience. It’s nice that a bank offers extra discounts for credit card purchases. The platform integrates seamlessly into the banking app such that users view these not as advertisements, but as general perks their banks offer. The existence of the platform increases logins and, with it, interaction with the bank’s services.
Bank customers enjoy savings on purchases with no corresponding negative exchange. In traditional advertising, the user loses her time and privacy to the advertiser, regardless of whether she wishes to be advertised to. In the case of targeted platforms such as Facebook, we know the controversies surrounding privacy, sensational posts, etc. In traditional platforms like print, TV, and radio, the consumer is forced to periodically sacrifice time to be talked at about how they should spend their money.
Unlike traditional platforms, the PI platform gives the consumer something: a discount. It is not a pop-up or a “click here to save 50% on your first order now.” It is a highly-targeted “offer” that requires a simple click to activate and no commitment to spend in the moment. The traditional exchange between the target of the advertisement and the advertiser is turned upside down. In this reversal, the target of the advertisement benefits with no loss. Furthermore, the customer’s purchasing data is anonymized and lives with their bank. This mean privacy concerns remain remote.
The PI platform provides highly visible return on ad spend (“ROAS”) unlike any platform currently in existence. The only platform that comes close in accurately and usefully providing ROAS information is Facebook. However, FB does this in an indirect way. It analyzes user behavior, clicks, views, etc. to estimate the return on ad spend. In contrast, CDLX has enough purchasing data to estimate, with a high degree of accuracy, the impact of an ad campaign on its platform. It can quite literally see the real-time change in consumer purchasing behavior when a new offer is launched on its platform. It can then feedback this information to the marketer who can more efficiently spend ad dollars. The PI data is also clear enough to help avoid the “Groupon effect” of low-quality, loss-leading sales and associated revenue cannibalization.
In addition to providing accurate ROAS metrics, CDLX has one of the highest nominal ROAS currently on the market. Although CDLX no longer provides this data directly and there are different relative ways to measure this number, it has reported ROAS as high as $30 return per $1 in marketing spend.
Cardlytics partners with FIs to provide them the PI platform. There are several key reasons why Cardlytics cannot be easily displaced:
- Integration is difficult and takes a long time
If you look back at the company’s history, you see that it takes several years for the platform to get up and running on a new FI and obtain high user engagement. The platform needs to be installed, fleshed out, and the habit formed with customers in order to be effective. Furthermore, the company needs to curate enough quality advertisers onto the platform in order to provide value to bank customers. This ramp up takes times, making a switch to a different PI provider difficult for FIs.
2. Banks and credit card companies cannot mimic Cardlytics’ business model
You might think that a credit card company can easily copy the CDLX platform and just provide this service itself. That’s not the case. Currently, banks and credit card companies are adversarial parties. They periodically negotiate credit card revenue split. Banks will not allow credit card companies to gain leverage in negotiations by opening up their customer purchasing data to the CC companies or ceding real estate on the bank website/app.
But, can’t the bank just provide this service in-house? No. Regulation and political/social scrutiny prevents banks from directly monetizing their own user’s purchasing data in this way. They need to outsource it to a third party, preventing them from dropping CDLX to provide the service themselves.
3. Network effects make new entry difficult
Currently, Cardlytics has a sales team that individually sells advertising campaigns to marketers. A new competitor would have to gain enough MAUs to even get into the conversation. However, in order to get those MAUs, they would need to come to an FI with quality advertisers ready to facilitate user engagement. It’s a chicken-and-egg problem that thoroughly blocks off access to new rivals.
CDLX drives revenue by 1) increasing number of bank customers who interact with the PI platform and 2) increasing the frequency with which the bank customers interact with the PI platform. I currently see major tailwinds (in addition to the strong positive trend pre-COVID) that will spur engagement with the platform, driving up MAUs and ARPUs over time.
- Habits form in recessions
We are only a few months into a recession. Eventually, people will start to tighten their belts. When that occurs, a simple source of friction-less savings on purchases can come from CDLX’s platform. Once the habit of checking your mobile app 2-3 times per month for savings takes hold, it can be hard to break. Any incremental increase in the habit will drive sticky MAUs. Furthermore, we will see current active users increasing engagement with the PI platform as saving takes up greater real estate in peoples’ minds.
2. Move toward a cashless society
Macro changes in the amount and frequency of cash vs. credit card usage only serves to increase engagement with the PI platform and value of the customers on it. I think this is self-explanatory.
3. Increased FI prioritization of the PI platform
As interest rates remain low and general economic conditions weaken, what used to be a nice side project receives a second look from important decision makers at FIs. If a better PI platform can drive higher engagement and quicker recovery for banks, they will become less restrictive on how far CDLX is allowed to take the user experience. If JPM allowed CDLX to simply put the offers box at the top of the app screen instead of at the bottom, think about the incremental MAUs gained and engagement increased.
4. Corporate marketing budget resets
The recession is a tailwind because it dislocates ad spend. In that process, CDLX is able to enter more conversations and highlight the extreme transparency and ROAS metrics its platform offers. As it is able to reach more marketers in the backdrop of resetting ad budgets, it provides a broader and better selection of offers for bank customers. Engagement would then increase, further improving CDLX’s value proposition to future marketers.
Currently, CDLX operates a simple UI and custom-sells campaigns via a sales force to marketers. CDLX brings in around $200m of revenue at this early stage. There is still a long runway for the company to milk its current strategy to generate high revenue growth. However, a direct sales model cannot be permanent for an ad platform to truly scale. I believe the company will be able to successfully implement a self-serve model in the long-run once its direct sales strategy peters out.
The CEO’s current goal is to roll-out UI changes that make the platform more user-friendly. The UI changes can not only facilitate user engagement, but create a platform organized by category, marketer spend, etc. to show offers across more verticals. From there, the goal is move on from a direct sales approach to a self-serve platform. It may be possible for your local coffee shop to bid for ads targeting users within a 5 mile radius who have visited any coffee shop in the last month. It may be possible for P&G to show you an offer for 5% off toothpaste if you visit CVS every 4 weeks and 3 weeks have just passed. The platform, given its access to such a unique set of data, has the capability to provide diverse, differentiated, and valuable capabilities at is matures.
CDLX’s valuation is attractive because it couples a strong probability of great returns with low chance of permanent capital impairment.
Outside of steady expensed R&D, the business requires no incremental capital to operate and has no debt. There is a ton of operating leverage in the model that can lead to high EBIT margins.
If we assume a 10-year ARPU of $3.50, EBIT margin of 25%, rise to 200m MAUs, and 10% discount rate on a 20x EBIT exit, we get an equity value around what CDLX is priced at currently.
I think this is a bearish scenario. CDLX’s ARPU numbers were near the high $2 range for its original FI partners and decreased as it integrated new partners. It is probable that the original cohort of users are generating near or above $3 ARPU by this point, a target which I think the general MAUs will easily reach as the platform matures and is integrated with the new FIs.
The nature of the platform, moats against competition, and the potential verticals it can access leads me to believe it’s possible to reach $10 ARPU in the bull long-run scenario and expand MAUs even farther. Furthermore, there are non-FI verticals CDLX’s purchase analytics capability may compliment which are not yet explored.
So, Cardlytics is a “quality growth company” that presents a solid risk-reward proposition. In the bear case, we are looking at market performing returns. In the base to bull cases, we are looking at a long-term double-digit compounder.
Update (August 2020):
It seems as though Apple is actively attempting to block Facebook from utilizing cross-user tracking via IDFA. This is particularly good for CDLX because the company operates its advertising with express permission of the hosting platform (FI partners). If IDFA significantly impairs Facebook’s ad personalization on a major mobile ecosystem, then CDLX will have the best ROAS and campaign analytics.