$CDLX - Getting Technical with Chase and AmEx

Another article on Clarity soon post NVDA, travelling so likely end of the week. For now we can get a bit in the weeds with tech at CDLX and the game theory underpinning that thesis.

I’ll likely also be on a podcast sometime soonish discussing CDLX. I think it’s attractive, misunderstood, blah blah heard it before. Stay tuned though!

Context:

If you aren’t aware there’s been a decent amount of debate recently on expert networks like Tegus and the like following Chase announcing their Chase Media Solutions. Even if you are aware, the transcripts alone aren’t always the full context, so hopefully this article is helpful in illuminating what Chase might be doing with Chase Media Solutions, what that means for CDLX, and how that can set up the stock rather well.

I generally try to think of what my “edge” might be in any debate, and here I believe it is on the technical/systems/incentive side. Essentially when I plug in all the available info my “game theory” estimates get a much different outcome than others, so I figured I’d highlight the logic and either influence the narrative or be told I’m wrong.

So we begin.

Chase Media Solutions:

For Chase’s overview click here and the initial announcement can be found here

Essentially Chase Media Solutions is branded rather identically to CDLX

“Chase Media Solutions campaigns come to life through Chase Offers, our customer-facing program. Offers are targeted using customers’ purchase history and published on Chase’s owned channels.”

Is basically what CDLX has been doing for Chase right, so I can understand the concerns.

Something to be established early on however is that this relationship has existed FAR before the official announcement earlier this year. Banks have requested a “self-serve” type platform wherein they can source their own offers since at least 2022. The way the relationship works is that the bank essentially submits the offer for CDLX to host within their Offer Placement System (OPS). The bank did the leg work to find the advertiser and do the targeting (if they so desire), CDLX acts as the infrastructure.

Naturally such a relationship likely leads to a lower share for CDLX - they are still being used, but they aren’t putting in as much effort. Terms aren’t disclosed, but that outcome would be logical. In theory banks could also architect things in such a way that their internal offers don’t require interaction with CDLX beyond fitting into the UI, so economics could be a risk.

There’s a couple mechanics here that need to be understood to accurately determine risk.

1. Offer targeting is actually quite hard. Typically there is a set budget, and you don’t want to cancel offers already in a users account. As a result you can’t just show every offer to everyone. You ask the advertiser for targeting parameters (new customers only/customers that shop at competitors/etc) then toss out some data science magic trying to predict what % of users will redeem, and thus how many users you need to show the offer to hit the target budget. Going back to an advertiser and saying “yea sorry we were 50% short of budget” doesn’t reflect well, and going over budget typically means you eat the loss

2. The data CDLX receives is not the holistic picture of any user. There isn’t a clear line, but it’s safe to say CDLX doesn’t know your name/sex/age/race/credit score/address/etc. They can probably guess based on transaction history decently well (Lululemon+Starbucks+Sephora, take a guess) but it’s reverse engineering instead of starting with the answers in front of you.

3. As a result, the current relationship functions something like as follows

   1. The bank sends CDLX transaction data with anonymous unique ID’s in a big expensive AWS instance or some equivalent

   2. CDLX determines target parameters with the advertiser and does data science magic to determine user targeting criteria

   3. CDLX asks the bank for approval (no competing promotions or “unsavory” ads)

   4. CDLX inputs user targeting criteria in their Offer Management System (OMS)

   5. The OMS sends the criteria to the OPS

   6. The OPS mixes the target parameters with a banks transaction data and determines which unique ID’s fit the criteria.

   7. A user logs in, which triggers an API call from the bank to the OPS asking for the offer list

   8. Optional Step: The bank overrides CDLX targeting and just asks for the full offer pool, then does the targeting themselves.

   9. Optional Step: The same as above but instead of overriding, they place some of their own offers into the pool.

   10. The bank presents the offers within their UI

       1. The ordering of offers in the UI is interesting. It can be random, sorted by CDLX, or sorted by the bank.

Those last few steps in #3 are rather key here when mixed with #1 and #2. Targeting is hard and banks often default to just letting CDLX handle it, but in theory they could be better than CDLX at targeting given more data. Historically when banks have tried handling targeting and ordering themselves, they just revert back to CDLX when performance degrades.

For what it’s worth I don’t think banks sucking at targeting is necessarily a long term guarantee, which is part of Chase’s rationale with acquiring Figg. Being fully reliant on an outside vendor for something you consider of strategic importance isn’t the best place to be, but also doesn’t mean you rip out the outside vendor arbitrarily, or that you can even outperform the outside vendor. It is entirely possible Chase gives up on their little experiment as others have before them, but I doubt it, hence the article.

Additionally, to be clear, the number of offers Chase has internally sourced is in the single digits nearly 2 years after acquisition, and they typically are rather small advertisers who I’d suspect are merchant relationships as part of Chase’s effort to provide merchant analytics as a combination within their business offering. If you’d like to jump down the rabbit hole of PoS and such you can have a go here. Some of the Figg team was spun into roles relating to these efforts and the overlap is quite similar with analyzing transaction data, etc, etc.

Thus the Chase Media Solutions ordeal isn’t a near term threat by any memes and the size of the team on LinkedIn is something like 90 individuals compared to ~70-80 when it was acquired in 2022. Not insignificant, but not super significant either if the goal is to have a fully fleshed out replacement.

In sum, the whole idea that bank transaction data is very valuable for marketing and other insights is playing out quite well, the key question is if CDLX still has a place to play in the offer landscape in 3-5 years if Chase has Chase Media Solutions? Another arguably more important question is what the CDLX/Chase relationship looks like over the next 3-5 years, it’s not like we’re priced for perfection here.

For an inkling of an idea, let’s cut to AmEx.

American Express: Why Now?

Something that doesn’t jive with a replacement thesis is the addition of AmEx as a CDLX partner. After all, AmEx had the most robust internal offers program and the cardholder perks are sort of the entire point of the company existing. If the optimal end state for Chase is replacing CDLX, why would the leader in the space turn around and put them on contract?

Clearly I believe CDLX won’t be replaced, so let’s try to understand the AmEx reasoning. I think there are a couple key variables:

1. CDLX is just better at turning transaction data into ad placement than anyone else. AmEx perhaps even behind Chase given Chase Figg acquistion, which 2 years later still isn’t producing many campaigns.

2. CDLX offers are complimentary to bank self sourced offers. CDLX primarily gets revenue from mass appeal everyday spend offers such as Starbucks, Chipotle, Walmart, etc. QSR’s, Airlines, large hotel chains, retailers, these are the largest CDLX clients and not what I’d imagine Chase Media Solutions goes after for a few reasons:

   1. If I’m correct that offers are more of a merchant benefit than a full fledged advertising platform, in which case these large everyday spend brands are not the target audience

   2. Chase would have to sell these brands on Chase exclusivity outperforming CDLX’s total network. Chase would need to more than double activation rates as well as spending years catching up on targeting and tech to attempt to match where CDLX is today, prior to the AmEx launch.

      1. Another way to get exclusivity is to self-fund the offers, similar to AmEx. This is not something that can be done en-masse.

   3. Chase cutting out CDLX removes the entire CDLX offer pool, which would put Chase at a disadvantage relative to other banks that stay on CDLX, unless Chase is able to poach many existing CDLX advertisers to exclusivity agreements.

   4. If banks focus on sourcing mid market merchant offers and self-fund targeted exclusivity agreements, they can achieve meaningful offer differentiation from other banks, while also keeping CDLX as a baseline of everyday spend offers.

   5. If you take a look through current AmEx offers, it’s somewhat similar to the picture I paint. The focus is more retailers/travel/experiences and more niche logos than the CDLX platform. The primary overlap is travel and the occasional niche retailer that puts a big budget through CDLX. My impression is the majority of CDLX offers standout quite nicely from AmEx and the other overlap can be combined for a more holistic logo experience and hopefully a higher likelihood of relevance.

   6. From the point above, it doesn’t make sense for Chase to be competitive with CDLX when they can do something similar to AmEx - Try to win some exclusive travel/retail/experience logos, mix in some merchant analytics benefits, and rely on the CDLX sales team for providing what they already provide. Why rebuild something that works well instead of augmenting?

   7. In sum, I believe AmEx (and thus likely Chase) view the offers as a net benefit to the entire platform. Think a traditional venn-diagram in terms of overlap instead of two circles fully overlapping. That logo perspective is in addition to an organization quite good at purchase data analytics and infrastructure.

3. If my understanding of the above two points is correct, and they are applicable to Chase, then AmEx likely had the view that Chase was going to be competing with AmEx more on the travel and niche retail side, with the benefit of CDLX augmentation. Take a look at the two case studies on the Chase Media Solutions site, Air Canada and a niche stove retailer selling $800 stoves. That’s textbook AmEx offer. AmEx didn’t want to fall behind Chase being augmented by CDLX, which is again likely an intentional push by Chase to overtake AmEx if at all possible. Why would Chase handicap themselves in the midst of a competitive push by cutting out CDLX? Building out a full replacement would push the schedule back by years.

   1. At an even higher level we can take a look at the cardholder benefits for AmEx vs Chase Sapphire. AmEx has their Uber partnership, while Chase Sapphire is stuck with a Lyft credit. This is likely a sore point for Chase given their size (holistically) relative to AmEx. Clearly cardholder benefits are becoming more of a relevant selling point with interchange pressure and Chase wants to be #1, yet isn’t. The logical conclusion for both parties is to accrue as much value to their platforms as possible. Replacement of CDLX by Chase does not accomplish that. Adding CDLX as AmEx accomplishes that.

AmEx details are of course sparse given the typical bank privacy expectations. The statement of work agreement details CDLX cannot use the AmEx brand name without permission, so the poor communication likely isn’t entirely their fault. CDLX is also ending up as a bit of a landscape for a competitive spat between two offer titans, so any specific workings are likely tight lipped for competitive reasons.

This of course sucks for public market investors and management has not communicated much of anything. We can only see a confusing ATM and convertible offering with little explanation. As a result the stock trading down on uncertainty isn’t surprising. As mentioned in my last article we’re left believing management is bad at capital markets with a good business, or snakes with a bad business. AmEx and Chase leaning in seems to indicate good business and intellectually it seems to jive, but definitely be open to alternative explanations. As a result it may be helpful to explore what the Chase Media Solutions team could be doing on the tech side.

Chase Media Solutions and Tech:

Below I’ll cover some things I think Chase can work on internally in a way that is highly mutually beneficial since the general perception of the tech side seems rather murky with many left in a state of “what else will they do but replace it”. My understanding of the architecture may not be perfect and it’s possible I’m wrong, but we’ll take a look.

A key thing to understand is that CDLX provides the data powering Chase and other banks offers, yet the banks are responsible for all presentation within the banks website/app. This leaves a LOT of space open for banks to optimize their offer platform on their side. Any changes would require significant effort given the multiple teams involved and stringent software development standards.

Some examples of what they might be working on:

1. CDLX has previously mentioned statement line offers/references to offers. This is something where the bank is almost solely responsible for that effort. The bank would need to determine and add in the UI element for that. They would work with CDLX on any targeting criteria or necessary API adjustments, but the majority of the implementation and decision making is left to Chase. Analytics on testing that feature would be a dual effort between CDLX/Chase.

   1. This tech can also be used for more than just offers - say a car payment pops up and Chase wants to advertise a refinance.

2. Working with analytics related to Personally Identifiable Information is another possibility. Chase can see variables CDLX can’t, and work in tandem with CDLX to fine tune targeting and results without actually sharing the PII with CDLX. These efforts can also be used for Chase internal reasons like providing offers specifically to say Chase Sapphire cards to provide differentiation between consumer tiers in an effort to up sell.

   1. Once again this tech can be used for more than just offers. Better transaction data analysis could help Chase better cross promote mortgages/consumer lending/car loans/etc. This is not competitive with CDLX and CDLX could even assist given their expertise. Chase gets numerous benefits by being better here.

   2. Analytics is very very labor intensive. Following Chase is somewhat illuminating here. It takes years to get content into the cloud and able to be analyzed. It takes a lot of effort to clean and work through trillions of transactions and test hypotheses. Figg was acquired 2 years ago and they’re still only beta testing really what’s possible. They have not achieved any meaningful offer scale.

3. Chase in the end state wants their property to be shopping enabled. We can see this a bit with their foray into car purchasing. Building out some key relationships on the sales side is an obvious way to work towards that goal. Additionally getting better at understanding how to convert consumers into transactions is a key skill to develop while working towards that goal. This can be done without fully owning the platform and I’d imagine the shopping angle overlaps heavily with the core offer market I’d envision they target, niche online retailers and travel.

   1. For those types of offers, Chase as a shopping platform can make easy sense.

      1. Larger ticket so more likely to be financed which is a cross sell.

      2. Don’t have their own captive audiences so Chase can aggregate

      3. Chase as an aggregator is likely better than AMZN margin wise

         1. Chase cares more about cross sell and consumer engagement than transaction margin, which is EXCELLENT for merchants. Traditional marketplaces or ad engines generate all of their value from the cut, necessitating lower margin on non-bank platforms.

      4. Chase as an attractive shopping platform is insane cross sell for other merchant services and analytics, which some of the Figg team was rolled into.

   2. There’s of course no guarantee lead gen turns into a retail platform. Google and Meta have tried similar things to middling success.

   3. Even if Chase is stuck in lead gen it can be quite accretive as they build out technology over time to better support additional services beyond just the offers themselves. CDLX providing the every day spend offers forms a base of consumer behavior and interaction. Chase wants to train users to interact with their app much more, and offers are a key part of that.

4. Chase travel is about the same size as their offers program, but both can fall within the same bucket. You land Air Canada as a Chase Media Solutions client? Now what if you interface with the travel team and set up a recurring offer with direct booking support to further decrease consumer friction? Architecting that project likely takes years on a bank timeline.

This list above is not exhaustive, simply an exploration of what may be in the works aside from a CDLX replacement, which doesn’t make sense for the reasons discussed.

There is still further confirmation to be done here that is in the works, but defaulting to an assumption of replacement doesn’t seem logical yet. If Chase Offers is indeed driving $8b of consumer spend per year while growing significantly, cutting that out seems a bit silly especially if competitive importance increases further. The recent Chase change to have CDLX higher up in the app/website seems to support this idea. Why show more customers the offer program more prominently just to lobotomize it in a year or two?

So how does it play out?

The Golden Path:

What I’d imagine happens is roughly as follows:

1. AmEx launches with CDLX starting late Q4 to limited accounts with a full rollout in Q1 2025 barring any headaches. CDLX handles the majority of offers while AmEx holds onto key exclusive relationships. In essence this is CDLX expanding MAU’s significantly to a set of MAU’s that are “trained” to interact with offers already at a card provider with offers as a differentiator.

2. Chase continues to iterate on the new ad server experience throughout 2024, beta testing some specific internal offer occasionally over that time frame as they better their backend technology.

3. Mid to late 2025 Chase internal offer technology is perhaps at the point it can support meaningful volume. The timing is a guesstimate from their current progress and trajectory. At such a point Chase tries landing competitive offers to AmEx and scrapping for exclusivity.

4. Exclusivity likely comes with subsidization from the bank side as otherwise merchants would not be incentivized to stay exclusive. Banks and CDLX continue testing how much of billings can be funneled to consumers, trading marginal monetary results for competitive positioning (as AmEx has done and will do given their lack of revenue share).

   1. In simple terms I’d guess a lack of revenue share largely accrues to the consumer incentive portion of billings, with CDLX benefitting in terms of volume.

5. Once Chase and AmEx start facing off more directly, perhaps in 2026 given AmEx would likely be firmly launched at that point with a ramped Chase internal team. At such a point differentiation likely comes more into focus vs building.

6. By 2026 CDLX hopefully has a full year of ramped AmEx, a Chase leaning in, and surely a WFC/BofA putting in at least minimal effort on the new UI. Adjusted Contribution not being 50% higher by then would be criminal (or I’m wrong about something).

7. At such a point cutting out CDLX would be a handicap in a competitive fight for offer relevancy and if AmEx or Chase defect, they’d have to bid for exclusivity. In theory the billings split at CDLX would be something like 40% Chase 40% AmEx 20% rest, so an exclusive network would be smaller than the remaining CDLX network, and less than half the size of what advertisers were running previously. Not the greatest selling point in my eyes unless Chase/AmEx subsidizes the spend.

8. CDLX by this point hopefully has some good iteration on the new UI, supporting visual customization at banks and working in tandem to allow for differentiation despite using the same backend. CDLX also hopefully has more scale to sell to advertisers, better analytics to ease the usage of the platform, and is consistently rolling promotions for large QSR’s/retailers/etc to drive the underlying demand to differentiated bank sourced offers and shopping.

9. The stock is $50+ entering 2027 and everyone lives happily ever after.

A pretty good outcome all things considered with an implied high growth rate and competitive lock-in as replacement would hinder competitive efforts of clients and shift their focus to re-creation of a platform instead of growth of a platform.

So naturally, how do we lose?

Warning Signs:

My process is typically to have a pre-set list of red flags that hopefully indicate business trajectory isn’t as expected prior to stock price declines. That isn’t always possible considering I often traffic in highly leveraged terminal value debate stocks where narrative is king, but hopefully it leads to more intellectual honesty.

Some things that might indicate being wrong:

1. Chase starts eating away CDLX advertisers. Do small retailers start working with Chase Media Solutions over CDLX? This can be tracked via the offer format as

   Swany407 Investment Research
   does for his subscribers. Thus far this hasn’t happened.

2. The AmEx contract doesn’t produce meaningful financial results for CDLX. Sadly if this happens the stock may simply tank, but if so it likely indicates terminal value is not very high and they get squeezed out by a lack of negotiating leverage. This seems unlikely given Chase’s move to give them more money and their comments about ongoing efforts to get lower bank share, but no guarantees.

3. Billings stagnates despite Chase’s effort to promote the platform. If Chase on the new UI with further offer pushing can’t cause growth this year, the stock working or not is entirely dependent upon AmEx economics. Thankfully for us the valuation isn’t demanding so this can still lead to positive outcomes, but would be a shame. Q2 and especially Q3 guidance coming in light on billings would be concerning and quite possible given the wide Q2 guide. It would imply customer churn and lack of results despite significant efforts from Chase to grow engagement.

4. Reversion of Chase user experiences. This is something that may be catchable prior to a stock drop. If Chase moves offers back down in the website/app that likely indicates it didn’t justify the space, in which case potentially nothing will and consumers just don’t like using a bank website/app as shopping. This would be puzzling considering they do it for travel indicating the issue is one of “training” which Chase is attempting, but no guarantee of course that it works.

5. Chase and/or AmEx start prioritizing Chase offers over CDLX offers to a significant degree. In theory they could have the typical 3 offer card highlight on the offer section all be internal offers. This seems silly as the goal SHOULD be to put up the 3 most relevant offers regardless of source, so any replacement likely indicates CDLX is viewed as a 2nd class citizen and the priority is internal offers growing, not the platform as a whole growing. This would also likely be catchable prior to any financial result impact.

6. Some other things I may think of later and tweet about.

Overall it’s a very dynamic situation with low info. This can cause rapid shifts in narrative/opinion for reasons I cannot fully list out, such as surprise capital market activities with dubious motives. The information economy here is quite tainted by the lack of communication and confusing mix of actions relative to implied outcomes (why capital raise if AmEx is relevant, why would AmEx sign if not relevant, why would Chase given partner share if not relevant, etc)

Fin:

Hopefully my future coverage is from a place with a bit more of a basis instead of ranting on technological possibilities and incentives, but anything is possible with this business. All taken together it does seem insanely cheap, so we’ll see what the lesson ends up being with regards to that.

Clarity Q1 results will be announced when I’m back home early next week, until then, cheers.