2019 in the mirror: Building Sakam
In late 2017, I rolled out the details of how The DApact ambitioned to rewire microfinance stakeholders after the early pilots in Cambodia were completed. In a nutshell, bringing about end-to-end matching and payments between loan seekers and impact-finance funds (MIVs), using traditional field lenders (MFIs) as non-custodial brokers. Let’s just draw it.
Here is where we left 2018:
✔️ Messenger-driven lead sourcing app beta release
✔️ Marketplace app alpha release
✔️ Lead sourcing PoC (pilot with pawnshops in Cambodia)
✔️ Realistic MIVs x borrowers stable-crypto onboarding PoC (demo with Wyre & Dai)
2018 was the year we moved from ideation & product development stage to Proof-of-Concept stage. Rewiring microfinance is quite an ambition, and by the end of the year we decided to breakdown our project into phases. PHASE 1 would consist in building a lending matchmaking marketplace, while PHASE 2 will be the end-to-end transactional channel.
In other words, we decided to pause blockchain developments and focus on our existing apps for leads prescreening & matching. Bringing this marketplace to life was already quite a challenge, but beyond that, we were hitting our heads really hard on the legal aspects.
PHASE 1 is a marketplace in the likes of AirBnb, eBay, Grab, with workable legal aspects.
PHASE 2 is an absolute legal mess in various aspects — loan portfolio ownership, licensing, crypto-channel legality.
The DApact had attracted quite some media interest and was labelled a “blockchain project”, which was counterproductive because of the whole ICO mess and regulators legitimately starting to freak out. So we decided to spinoff the marketplace as a stand-alone project. Introducing Sakam!
Sakam is the brand name we created to id PHASE 1 — marketplace. Our one objective for 2019 was to make Sakam a tool at work, building a sell-side and buy-side in the marketplace.
Here is where we’re leaving 2019:
✔️ Lead sourcing app V1.x LIVE
✔️ Marketplace app V1.x LIVE
✔️ 9 signed MFI partners
✔️ Borrowers / MFIs matchmaking (PHASE 1 — LIVE)
✔️ EoI from MIV partners
✔️ We have a business model (finally!)
Our model is to source loan applicants online using our consumer-app, and dispatch these loan applicants to partnering MFIs through our lender-app. Together these apps form Sakam Marketplace.
Our value proposition to consumers is to optimize credit value-for-money (or simply optimize their chance to be approved for a loan) by distributing their request & profile to several lenders at the same time, which drastically increases their negotiation power.
Our value proposition to MFIs is to provide unmatched cost- and sales- efficiency throughout their origination process, by sourcing, processing and prescreening loan request through a full-blown AI, no human-agent process.
Let’s dub it “the 70-cube problem” :
70% of a typical MFI cost is payroll. 70% of payroll is field staff. 70% of field staff’s work is low added-value tasks.
Field officer’s low added-value tasks entangle going back-and-forth with potential loan applicants to gather KYC and biodata, documentation, form filling and digitalization, first-stage screening, etc.
Hacking loan origination
We solve the “70-cube problem” by programming a very dull credit officer, but one that works 24/7 and can handle 1000’s loan requests simultaneously, and instantly push them to a bank’s back office. It’s dull and as such it’s just extremely good at performing the dull part of credit officers job.
And no joke, our consumer-facing app is a gem of Human Centered Design (and subclass Culture Centered Design), as well as a tentacular combination of AI subfields: Computer Vision & OCR (trained to recognized single-image local documents with > 95% success), Speech-to-Text + NLP to handle complex biodata input (AFAIK a premiere for Khmer language), and above all a conversational app that’s robust enough to deal with the kind of otherworldly usage you (wouldn’t-even-remotely) expect with tech-unsavvy and generally little educated people.
Getting there has been a three years long process of designing, coding, testing, doing it again. I personally got to learn to read Khmer by dint of going through endless user’s chatbot conversations and programming OCR and Speech-to-Text custom layers. A loan application is very much comparable to a complex gamebook, a decision tree mapping that must adapt to various kind of loan services, biodata and documentation. This now translates into 70K lines of proprietary code.
The consumer app must also fit in with partnering MFI’s policies, so that all request can be covered by at least one partner in terms of service + requirement. As to limit competition in between partners, we had to cherrypick MFIs with disparate financial inclusion focus, while making sure we had sufficient geographic coverage. Reaching out to big banks is not an easy job for a small startup — not to mention that we seriously need a decent business dev in this team — and only with the help of industry influencers (🙏 Dina Pons) could we get a foot on the ladder.
As it turned out we’re eventually working with the biggest and best-in-class MFIs in the country.
Onboarded lenders get access to Sakam lender app, where they can explore (1) projects seeking fundings, (2) project-holder biodata and (3) confirmed financial guarantees. Project/lender matching is a staged process that eventually whitelists one lender for exclusive access to project holder’s KYC data and contact lines. Sakam’s mission really stops at this stage — when a project is booked for underwriting — although the app also includes CRM features to notify important events, streamline contact scheduling, obtain GPS coordinates, etc.
If a project is successfully funded, Sakam draws a 1% commission on the requested loan amount. So, by paying just a 1% fee, a lender gets a qualified, prescreened customer with CRM facilitation. On the other hand it has saved 100% of its advertising cost and 70% marginal payroll cost (i.e. field staff hours). Oh, and we never ever charge any other fee to lenders, be it on setup or as a membership fee.
In the event a project underwriting closes on a reject call, the project is brought back into the global pool where it gets another chance to be funded, with a mention on previous underwriting findings.
Indeed, MFIs are expected to cooperate in between them within Sakam by disclosing reject reasons, so as to avoid multi-time assessment of insolvent projects (due to overindebtedness, etc.). Enhanced cooperation between members really is something we want to explore in the coming months. Notably, we are working on portfolio cross-financing, where a commercial bank will offer limited credit lines to MFIs financing loans through Sakam. Up to 70% of every loan principal could be financed for a mere 10% interest rate — a bargain for small to medium MFIs.
I think getting to unite a nation’s main financial services providers on a single marketplace is some kind of world premiere, and immensely beneficial to end-consumers. Especially in microfinance, which is characterized by the diversity of recipients’ needs & profiles and little possibilities/abilities to benchmark their market. For a reminder, there are 75+ MFIs in Cambodia and 57 banks — that’s one for every 25,000 households — and blackmarket lending probably accounts for over 50% of disbursed loans units. Depending on who hooks you up, the interest to be paid would go from as low as 18% to x00% — and yes, 18% is an as-low-as, this is what we are trying to deal with at The DApact.
So what’s our next step? Well, look at this:
Although the presentation is the same, profiles in Sakam bear qualities that Kiva profiles don’t (see Kiva business model FAQ to understand why):
- Unaffiliated: they do not belong commercially to any specific lender
- Live: they correspond to existing, current needs
- Actionable: there’s an open two-ways electronic communication channels
Social investor’s impact is essentially conscripted by the lack of direct access . trust / contact with their ultimate beneficiaries.
Sakam aims to leverage the digitalization of traditionally remote population to bridge the gap between mission-driven lenders like Kiva and their ultimate beneficiaries.
As cash will become digital, traceable and “smart-contract’d”, we believe there is a window for direct subsidies of loan seeker’s whose financial project fits in impactor’s program.
This is the scheme we’re eyeing with our friends Good Return, the Australian Kiva. Considering a game where Good Return could directly subsidize due interests on a microentrepreneur’s loan provided by Chamroeun MFI, there would be a win-win-win situation:
- Improved social impact and marketing power for Good Return
- Improved loan value-for-money for microentrepreneur
- Improved competitiveness (service rates) and reduced risk for Chamroeun (e.g. subsidy could be discounted from the last installment only)
As we are dealing with unbanked populations, such system involves cryptocurrency value transfer, which poses a number of UX and legal challenges. But Cambodia might be a unique chance to iterate over this idea. Three options come to mind:
Dai (stablecoin) is an Ethereum token we used to demo realistic MIVs x borrowers stable-crypto onboarding in 2018. Our legacy blockchain stack sits atop Ethereum and it would make sense to build on our existing stack. But turning Dais into fiat is virtually impossible without a crypto know-how, and there isn’t any sign it would be any legal either. We now tend to dismiss Dai as a viable option.
Our consumer app is built over Facebook’s tech, and Libra comes as a natural cryptocurrency candidate. Libra’s main advantage is it may literally abstract all UX challenges of holding cryptos. More importantly, if it was to be whitelisted, Libra would come with a full regulatory frame, bringing about viable guarantees for companies such as ours. Libra’s source code is already available on Github, and *advertised* introduction date is second half of 2020.
A world first, the National Bank of Cambodia is about to issue its enterprise blockchain CBDC. Bakong has been in the making for a while and the NBC has been real serious about it, with plans to gradually built it as a cross-border payments channel too. CBDC is an unmatchable combination of good UX + on/off-ramp + legal feasibility — Sakam could just simply drive subsidies to borrowers on condition they own a Bakong wallet. Bakong’s underlying tech supports some sort of Smart Contracts, but a crypto on-ramp seems like a long shot.
Overall, the prospect of streamlining foreign impact capital through the blockchain to subsidize Cambodian’s microprojects is thrilling, but it is highly contingent on a pre-existing, regulator-approved on/off ramp.
While we do keep an eye on the exciting developments 2020 could surface, we will primarily focus on developing Sakam marketplace in the coming months.
Here are our plans for 2020:
- Involve more MFIs and MIVs in Sakam
- Portfolio cross-financing with our partnering commercial bank (B2B scheme)
- Myanmar expansion feasibility study
- Workout the legals for PHASE 2
- Pilot PHASE 2?