These are chat archives for fairydust-agency/project-wasl-grants

15th
May 2016
AshrafNassef
@AshrafNassef
May 15 2016 09:03
@ninabreznik Your understanding of WASL is correct.
WASL will derive trustworthiness scoring from social data. But social data not only derived from the interaction with the platform but also from other social media channels like Facebook, tweeter, LinkedIn, etc.
I know it may not be clear how. It is even not clear for me. But some banks are already making credit decisions based on the big data analysis and scoring of the applicant social data.
Please have a look to this example: https://www.kreditech.com/what-we-do/
We should not start with something very complicated but we can build it over time. Our first patch of CMs/supporters will be people and organizations that we should already know. It will start with a trust network rather than depending 100% on the scoring.
I think of partnering with universities and research centers to help us in developing this scoring system. I think it is very much possible. What do you think?
@ninabreznik Regarding the problem:
Please remember that WASL will only provide its services to the New-Comers who are legally eligible for work. The ones who have the official documents that allows them to work.
WASL will exist to specifically serve newcomers with no credit history (we will build it for them), are not integrated in the local community (we will link them with the supporters/change makers and help both build the relationship for a purpose)
@ninabreznik How does WASL rate trustworthiness of change-makers/supporters?
Supporters are two types:
Individuals: could be scored through referral system (WASL shall trust someone referred to it by you for example) and other social parameters/criteria. To be developed.
Organizations: also could be scored through referral and reputation and history of successful support to new comers.
Again, We should not start with something very complicated but we can build it over time. Our first patch of CMs/supporters will be people and organizations that we should already know. It will start with a trust network rather than depending 100% on the scoring.
@ninabreznik I love the idea of the business app.
WASL platform should be designed from day one to be able to integrate / add new features all the time and grow to serve many people and many purposes as they emerge.
The only point is to set priorities. What are the main basic features to start with to launch the platform and what to be added at a later stage after to test the basic features with the beneficiaries?
What do you think of this?
Did I answer your questions? Or I created more confusion?
AshrafNassef
@AshrafNassef
May 15 2016 09:13
On a separate note and not to create any misunderstanding, please change the fairy dust agency to WASL instead of AKDN?
I am working for AKDN but AKDN has nothing to do with this project.
By the way, WASL is an Arabic word means connect / union / bring together.
Do you like it as a name to this project?

Otherwise - not all of our issues were adressed regarding the project in general...
As already mentioned, we see the following issues:

  1. ISSUE - lots of refugees in short time that need to be processed,
  2. ISSUE - refugees often start with just a fingerprint and picture (lost passports etc.)
  3. ISSUE - often have no credit history that can be accessed (no data),
  4. ISSUE - no existing longterm relationships between WASL and local CMs,
  5. ISSUE - CMs and NCs don't know each other, not from the same country, different language, culture.

I will try to elaborate on the "issues" we identified in context of your answers.

WASL Project

Please remember that WASL will only provide its services to the New-Comers who are legally eligible for work. The ones who have the official documents that allows them to work.

Yes, that's what we assumed already. According to what we know it approximately takes 1 year before new comers get a work permit, but they can already work in similar ways that students can work before they have a work permit to work full time. Students can register to work self employed, so this might be an option for new comers too.

    1. We strongly recommend to partner up with lawyers to get detailed insights in what is possible and what isnt and invite them to join the project as soon as possible

We should not start with something very complicated but we can build it over time.
WASL platform should be designed from day one to be able to integrate / add new features all the time and grow to serve many people and many purposes as they emerge.
The only point is to set priorities. What are the main basic features to start with to launch the platform and what to be added at a later stage after to test the basic features with the beneficiaries?
What do you think of this?

Yes, we totally agree to follow an agile and lean approach of developing the WASL product/service over time, but in order to maximize the benefits of each iteration it is important, that there is guidance from a long term vision/concept that is as detailed as possible before starting development. Creating a detailed concept is relatively cheap and the benefits from the aid provided by it when making decisions about the priority of features to develop in each iteration is huge.
We need a very detailed concept, thus we need to break down a very detailed vision of the future into features of the platform in order to be able to prioritize features that steer the WASL project in the right direction in each iteration.

WASL will exist to specifically serve newcomers with no credit history

Yes, this is already our understanding, see [Issue 3]

(we will build it for them),

Yes, this is what we think is vital.

  • Q1: How to build a credit history for them?

are not integrated in the local community

Yes, this is already our understanding, see [Issue 5]

(we will link them with the supporters/change makers and help both build the relationship for a purpose)

Yes, this is what we think is vital.

  • Q2: How to choose CMs and how to aid them in building a relationship with NCs?

Regarding Q2 (Question 2)

Our first patch of CMs/supporters will be people and organizations that we should already know. It will start with a trust network...
  • Q3a: How well do you really know potential CMs/supporters?
    It's always possible to create financial and other incentives to compensate, but
    in order to develop the WASL platform it's important to create a solid concept about
    those incentives now. So Question 3 can be replaced with:
  • Q3b: How to set the right monetary and/or other incentives?

Individuals: could be scored through referral system (WASL shall trust someone referred to it by you for example)

That "outsources" the problem [Issue 4] from WASL to find many partners, to WASL finding a few partners that will find more partners.
We believe, even though it might be possible to find CMs that are trustworthy, whether they will disappoint or not is strongly connected to transparency that will slowly replace trust based on

  • hopes and goodwill (=guts feeling)

with a

  • solid reputation score based on measurements.

and other social parameters/criteria. To be developed.

This is what should be developed now - even though it can be implemented later, we believe it's vital to work out a fully featured concept that allows to derive the MVP (Minimum Viable Product) that lays the best possible foundation on which to grow.

Organizations: also could be scored through referral and reputation

So finding initial Change Makers (Individuals and Organizations) is basically a random shot in the dark guided but nothing else but human guts, which seems ok to us if we can develop a solid concept about how to ...

and history of successful support to new comers.

... continuously measure and evaluate the relationship with Change Makers.

I think of partnering with universities and research centers to help us in developing this scoring system. I think it is very much possible.
What do you think?

We think universities and research centers if they are willing to help, are good initial candidates for CMs and can also provide a lot of aid when developing a scoring system.
BUT: we also think, that the critical problem is not to build a scoring system - this is actually the easy part and the knowledge how to do that is relatively easy accessible, see:

The three main problem that we see are:

One of the reasons widely mentioned by global financial institutions reporting financial inclusion issues is related to financial identity. Poor population in the developing world often doesn’t have a formal financial history, or even financial identity (not mentioning that some don’t even have identity documents).

  1. Financial Identity - Q4: How to build a reliable digital identity for new comers to which a credit score can be assigned?
    • This might be relatively easy given, that opening a bank account, registering a business, showing registration papers to CMs in a face2face meeting will make it practically impossible for newcomers to create "fake identities".
    • If not enough, additional methods can be utilized, e.g. SIM cards,biometrics,etc...
    • NCs (=newcomers) in germany can easily get a bank account now.
  1. Q5: How to provide relevant skills and education to new comers in order to maximize success rate by building a solid concept for active transformation in addition to just passive measuring of scores?
    • Pure passive measurement and evaluation or data might work, but doesnt tap into the potential of providing quality business education to newcomers that will increase chances of success.
    • Our suggestion would be to create a lot of transparency, so that new comers can easily learn from each other
  1. Financial History - Q6: How to get relevant solid high quality data continuously in order to feed them into whatever scoring system we might use?

WASL will derive trustworthiness scoring from social data.
But social data not only derived from the interaction with the platform but also from other social media channels like Facebook, tweeter, LinkedIn, etc.
I know it may not be clear how. It is even not clear for me. But some banks are already making credit decisions based on the big data analysis and scoring of the applicant social data.

We are familiar with this practice. Kreditech is one of many startups that ride that wave. We have no doubt, that the techniques used by Kreditech are able to provide very meaningful insights given the underlying data is of high quality.
Our recommendation is to be careful and not believe the hype.
Kreditech is only 4 years old and other similar companies face the same problems.

Some background information

  • This approach becomes more widespread, see here
  • They officially started out in 2012, but actually existed before under a different name
  • Kreditech is one of the pioneers of the approach you mention. They do not have access to traditional credit history, which is why they started lending their investors money to people in order to gather financial history data. This generates a very constrained sample data set. Critics say it's hardly possible to derive any meaningful insights from it, giving little credit to attain unbiased labeled data set.
  • Recently (October 2015) they still remain unprofitable
    • Source1
    • Source2
      • €50 million – revenue run rate
      • €120 million – issuing run rate
      • 2 countries ramp-up (Romania and Brazil)
      • 9 countries live (Poland, Spain, Czech Republic, Russia, Mexico, Australia, Peru, Kazakhstan), further plans:
    • Dominikanische Republik, Rumänien, Argentinien, Peru, Brasilien
      • 200 employees of 40 nationalities, 40% of which are women.
  • Despite their "modern approach", they:

    • let only ~15% of applicants pass their test
    • and still, there high interest rates are considered usury
      • In many cases more then 100% p.a. interest and in the most extreme case more than 3000% of interest
    • Kreditech is not allowed to operate in germany (-> Source)
  • Please read criticism https://en.wikipedia.org/wiki/Kreditech#cite_note-3

  • The former CTO is now CEO and we can recommend this presentation of him:

    • https://www.youtube.com/watch?v=Z57UVGxt_6E
    • He gives some insights in their shady evaluation system, which includes KPI's like:
      • area of living to derive their monthly costs for paying a room
      • willingness: filling forms + typing patterns => is it suspicious?
      • linkedin profile for assuming average income for whatever they list
      • compare facebook talk about location with what comes from other sources
      • they don’t require any external credit bureau data to generate their credit ratings. Instead their algorithm is based purely on globally available data sources including social networks and e-commerce records.
      • visited web sites
      • used font when sending pdf applications
        • certain fonts are only used by online casino/poker/... platforms
      • Please also see here
    • how to compare constrained data sample
Nina Breznik
@ninabreznik
May 15 2016 17:28

So, to sum up:
WASL's key value proposition is core competency is to develop and provide a reliable credit score to its potential investors/creditors/partners.

We believe that modern statistics and algorithms can be used to create a high quality credit score, but for that we need high quality data. We think the reliable data needs to be the essential part of WASL's core value proposition.

Therefore, because reliable data about newcomers doesn't exist, we have to generate it.

How? By providing them with business software in which all data is transparent.

So, we think, even for MVP, we would need to build the platform around this business software...

Nina Breznik
@ninabreznik
May 15 2016 17:35

Our suggestion is to start with a TODO app where you can:

  • share Todos/Tasks/Pins and
  • issue invoices.
    And then we add features on top of it.

What do you think would be the most essential features to start with? And to understand the whole project more in details it would be great to know what is the budget, who are the stakeholders involved, what are the interests of people and organisation(s) behind it? Is it just your personal pet project? What else in your opinion would help us understand the project better?

Alexander Praetorius
@serapath
May 15 2016 18:12
Alexander Praetorius
@serapath
May 15 2016 18:41

Our core idea for the app is to create a TODO App similar to any.do

  • see https://www.youtube.com/watch?v=gPpHp4Yfs_M

    And implement PDM or CPM behind the scenes

  • see https://www.youtube.com/watch?v=jmCc5VIMOro
    This can be used to create business plans and financial projections, refine them and later invoice by delegating an invoice todo. It's easy to use as any.do and continuously turns financial projections and the business plan into operations and actual financial results that can be compared to what was planned
    • WASL can use it to transparently derive credit scores
    • Change Makers and Investors can use it to dig into details in case the credit scores are not enough for them
    • Newcomers can use this transparency to build trust and even learn from each other to maximize chances for success
Alexander Praetorius
@serapath
May 15 2016 19:23
Through "sharing tasks" (including invoices) within the structure described by the videos above, it will be possible to model business processes similar to how large organizations do it using languages like BPMN 2.0 and measuring to a fine grained degree whats going on and how to optimize.
Alexander Praetorius
@serapath
May 15 2016 21:29
Example for making an "app demo": https://popapp.in/
just click and press play to see how it works - we will try to make it similar
Alexander Praetorius
@serapath
May 15 2016 21:36
(One more thing: the todo app above that works with sub & subsub... tasks behind the scenes to have network diagram functionality built in, will of course be a so called "work break down structure" of the Business Model Canvas and Business Plan into lots of actionable tasks)
Nina Breznik
@ninabreznik
May 15 2016 21:58
13219600_10154696736961416_775335090_n.jpg
MVP sketch