15+ Platform investor - happy to answer questions

[fintechjunkie]

Curious about the questions this audience is interested in having answered by a serial investor in the space. Let me know and I'd be happy to share some thoughts:

www.fintechjunkie.com

In particular, this audience might be interested in my white paper that details the last 10 year history of the personal loans space and what's likely to come next:

http://fintechjunkie.com/2015/04/10/...eing-released/

Send questions!!!!

[platformlending]

Will definitely give this document a read!

[fintechjunkie]

New Blog Post available

If you're interested...my newest BLOG post is up. Another installment about mistakes made in pitches.

www.fintechjunkie.com

Enjoy!

[Investor]

Your piece was probably the most thoughtful piece on the industry I've seen so thanks for putting it together, will have some questions I'll post later after I reread and think about things.

[fintechjunkie]

Happy to answer questions. Working on a SMB lending piece to be released at Money2020. Hope to help however I can!

www.fintechjunkie.com

[Investor]

Great, it was a fascinating read which I really enjoyed and it also sparked a bunch of questions some of which are below.

1. How do you think about the accuracy of automated underwriting relative to traditional underwriting. These platforms are cheaper and easier for consumers but I wonder if to some extent that is because they aren't pricing risk correctly given they haven't been through multiple cycles. Obviously some platforms were around in 08 but one can only wonder if their credit quality has eroded as those players grew and they have been operating at what has generally been a healthy last 6 years. These guys avoid a lot of regulatory costs, branch office costs, etc. but I would presume those physical offices provide some benefit in customer acquisition too so would be curious about how big of a cost savings you actually thought they had as I'd guess customer acquisition costs will be a lot higher in 3 years for the space (will be exciting to see how Credit Karma turns out and I was always impressed with them). Just would guess that some of the smaller guys aren't pricing correctly but curious about your thoughts and how widespread this may be.

2. LaaS is a very exciting concept. Why do you think banks are slow to adopt it and wouldn't the biggest players (prosper, Lending club, etc.) dominate this market? If I was a bank I wouldn't want to be taking credit risk on a new entrant with an unproven underwriting record. Do you think Prosper is disadvantaged relative to someone like Lending Club because they only do consumer loans? Just wondering if banks would want one LaaS partner and Prosper isn't a 1 stop shop then (granted I saw they partnered with Ondeck in some form).

3. How is direct mail a scale game? Don't want you to go into anything proprietary about any of your companies but credit card companies have been doing direct mail for years and it seems like just about any finance tech company has managed to wrangle enough VC money to be able to hire analytic and creative marketing talent.

Thanks would be very curious on any thoughts you could share.

[Investor]

And 2 additional questions I just thought of, and apologies if these come off as offensive/questioning the business models, but I think they are valid questions.

4. There was an article that mentioned how everyone thinks they have unparalleled expertise in data science and think that the banking industry doesn't have sophisticated models because their technology wasn't capable. Then it discussed how all these fintech startups think they can use their algorithms, social media data, etc. and disrupt the credit card business, payday lending, etc. but in reality banks have already tested that and while it worked great in testing it failed over a market cycle. Do you have any thoughts on this? I'm presuming their is better data available now for data scientists than previously but would be interested on your thoughts.

5. Do you think these lending platforms can actually become mainstream? I guess this relates to the above question about how much this underwriting actually works. I went through a Lending Club presentation where they said that banks have operating expenses of 5%-7% of their origination volumes and lending club is 2% (if they stopped growing) so they had a material cost advantage. That cost advantage was partially due from they had some costs the banks didn't bear at all (no branches, no regulatory costs with collecting and maintaining deposits, etc.) and they had some costs that were less than banks as certain things were automated with technology (underwriting, servicing, etc.). If you say Lending Club has a 3.5% cost advantage and let's speculate that is attributed to each of the 2 categories. The banks branch costs, regulatory costs, etc. I'm presuming are largely fixed so in that case on incremental business for the bank LC has a 1.75% cost advantage through using their technology. Is that enough of a cost difference to disrupt the space? The other interesting observation is it seems like these major platforms are, at least for now, lending largely for debt consolidation and taking share vs credit cards which is almost an unfair fight because credit cards aren't really doing risk based pricing. I'm just trying to think about how they match up head to head vs banks on comparable borrowers, better understand these cost advantages, and try to better understand the accuracy of the underwriting relative to humans. If a marketplace has a 3% cost advantage but loans default at xx% higher throughout the cycle then there can be an interesting discussion about the ability of these platforms to scale outside of niche use cases.

I'm actually getting excited about the fintech industry but also happen to enjoy debating things and in this case I couldn't pass up a "Send questions!!!" comment from someone who knows the space so well. Will spend a bunch more time researching this as it looks interesting.

[fintechjunkie]

Good questions. Here are a few quick thoughts:

1) Underwriting is a tricky animal. It's not about "automated" vs "traditional" underwriting. It's about the quality of the models. If a credit policy relies only on models that are grounded to specific outcomes then the credit performance will be subject to variances based on economic conditions that didn't exist in the modeled sample. If a credit policy relies on traditional underwriting policies, then it doesn't understand the nuances of channel performance, positive/adverse selection, pricing, etc. The best of both worlds can exist --- grounded predictive models based on data with important overlays that account for variables that identify toxic customers in an economic downturn (i.e. - DTI ratios). The best companies have great policies....but there are some pretty crappy companies out there as well so buyer beware.

2) Banks are slowly turning towards the originations platforms and seeking out their services. This will just take time because many Banks operate at a very slow pace and are adverse to trying anything that's new without many reference clients that they can use as proof to cover their butts within their own organization.

3) Direct mail is a scale game because costs come down as volume goes up. There are companies that are reinventing this equation through API integrations (LOB is a great company in the space) but fundamentally direct mail is a competency that requires many cycles around the track to get right (i.e. - response and risk models tuned to product and creative results).

Hope this helps! I'll answer your other questions in the next post.

www.fintechjunkie.com

[fintechjunkie]

No offense taken --- you're entitled to your opinions:

4) Machine learning modeling techniques are proving to provide significant lift over traditional models but this isn't secret sauce that can't be replicated by the big banks. The trick is being able to use "big data" (I hate this term) with "machine learning" with "economic overlays" in a "compliant manner". Getting all of these dimensions right is tricky (especially being on the right side of Reg B) so even the best companies with the best ideas and the best data scientists aren't guaranteed to be able to use their skills if they don't have a proper understanding and appreciation for the regulatory and economic environments.

5) Yes - these platforms can go mainstream. It's only partially about cost savings. It's also about competencies, focus, customer experience, etc. When put together properly a specialist can overcome their cost of funds/leverage disadvantage and create a very compelling business with great returns.

Keep them coming!

www.fintechjunkie.com