The Other End of Sunset

Thursday, May 10, 2012

This End of Sunset

Hello OtherEnders...

I know it's been forever since I've written here, and I guess I want to celebrate the passing of TheOtherEndofSunset.  I've written some funny things (I think, anyway), and a load of sad stuff.  But it's all been honest, and I've tried (very) hard not to talk about business here.

But this phase of my writing is over. At least for now.

I'm writing a new blog -- entirely business focused -- over @ Forbes.  I hope you will like it.  You can find it at -- pretty great branding, eh?

The Forbes blog has my political viewpoint, my humor (whether you like it or not!), and song lyrics. But it won't talk about my personal life (or at least I don't think it will).

If you like my writing, I'd love it if you'd read the stuff I write on Forbes.  I'm having a good time writing it, and I haven't written anything that makes anyone cry!

Hope to see you over there, comments welcomed as always...


Friday, October 14, 2011

French, and English, at the same time

Hello my OtherEnders.

It's been a while since I've felt like writing.

I can't get away

I've done some writing for work, and I've really meant what I said. I built a company that can -- and will -- save the underbanked billions of dollars. I plan on transforming an entire industry. Every minute I spend thinking about it, every new hire we make, every new product we push… they all make me more convinced that ZestCash can change the world.

But there is another part of me that writes. It's sad, it's forlorn. (And, apparently, it's redundant).

I never knew it was part of me. It first raised its ugly head when JR got sick.

That side of me is driving right now.

It's understood
that Hollywood
sells Californication
--Red Hot Chili Peppers

I described Jeanne's death here. She passed on in the afternoon of June 23rd, 2006.

// Side note: "Passed on" is such a silly way to say "died", don't you think? You get passed on an expressway, not in life, nor in death. //

When Jeanne died, I was holding her hand. I couldn't feel the minute she died, like in the movies. Her hand was cold, slightly blue, and still, even before she died. The nurse told me when she died. Just about 1:30 in the afternoon.

I didn't feel anything at first after she died. Just … quiet. Still. Not good quiet, like in a library. Bad quiet, like the stillness between a heartbeat and the next when you have hurt someone, badly, and they find out.

Then I took a handful of Ativan and went to sleep. Everyone left the next day, but our two rescue dogs were still with me. Minnie, the Dalmatian, and Tyrone, the Brown-Dog-of-Mixed-Parentage.

Back on the chain gang
--The Pretenders

Almost two years ago, Minnie started walking with her head down, below her shoulders, and dragging a foot slightly. We thought she had a back problem, so we gave her pain killers, and assumed she was just getting a bit arthritic.

Then it got worse. Much worse. Suddenly.

We did X-ray after X-ray after CAT scan after MRI … And nothing made sense. Until the doctor had the idea of scanning Minnie's brain.

Revealing a large tumor that was killing her.

// Side note: Ever since we've been a couple, SL has tucked in both dogs every night. She has a particular set of words she starts with, and ends with, and a cadence that both dogs recognized. It's a thing of beauty, light splayed across a wall from a prism in the window, the first day of spring. You should envy me for getting to hear it. //

A few weeks after we found her tumor, Minnie lay in bed next to me, with her eyes partly closed listening to SL giving her a final tuck-in. The tuck-in ended just before the vet injected poison into her front paw, stopping her heart.

I knew when Minnie died. I could feel her heart stop. And feel her postural muscles give way.

Tyrone missed Minnie a lot for a while, but then settled into his role as family guardian. He would sit between SL and anyone he didn't entirely trust. When she was pregnant, Tyrone stared balefully at SL's personal trainer every time he touched her. It was clear that he was watching.

When Elisabette was born, Tyrone changed his focus from just SL to cover our new pack member as well. He sat in the doorway to Elisabette's nursery when SL was feeding. Nobody was getting past him. His job was to protect everyone when I wasn't around, and he took that job seriously.

A few weeks ago, Tyrone started walking more slowly, his head below his shoulders. This time, we didn't think it was his spine. We found a large mass, about the size of a tennis ball, just behind his front leg.

He had cancer of the blood.

Every day, he'd get a bit slower, and have a bit more trouble getting up the stairs. But he still blocked the doorway if I wasn't there, and he still barked at the doorbell.

He wasn't going to let cancer stop him from doing his job.

Two days ago, he lay on a blanket, in our front yard, listening to SL give him his final tuck-in. He lifted his head and looked at her. As the drugs took over, he lowered his head, but didn't stop looking at her until the end.

I had my arms around him. I felt him die.

I don't like illusions
I can't see them clearly
-- Sick Puppies

// Side note: Yes, I perceive the irony of the song lyric I just wrote. It wasn't accidental. //

God, I hate cancer. It shouldn't exist. We should have cured it. Or made it a chronic but manageable disease.

We can bomb arbitrary countries back to the Stone Age, but we can't cure cancer?

It's possible we are spending our time and money on the wrong problems.

// Side note: I almost always write on long plane rides. Is it possible I don't like traveling because this part of me wants to come out? //

Sometimes a change
is not enough

I have mixed feelings about death. Odd, I know, but bear with me for a second.

I miss Jeanne. Every day. I miss Minnie. Every day. I miss Tyrone. Every minute.

For good, or for ill, the pain fades over time. It becomes less like a knife, and more like a sprain. It doesn't cause agony, but you don't especially want to stand on it.

I think that analogy stinks, but it's the best I have. Sorry.

The mixed feelings are that I got to be with Jeanne at her last conscious moment, and her last Earthly one.

I got to hold Minnie's head as she lost the ability to hold it herself whilst shedding her mortal coil.

I got to feel Tyrone's breath fading from his body. He wasn't scared. He was strong, and brave.

In my heart, I know he's chasing Minnie around now, happy as a clam to be young and healthy. But I also know he wishes he could stand between SL and the rest of the world.

// Side note: There's nobody in the window seat next to me. Which is good, since it lets me look out the window to hide my tears. Not so manly, am I, now? //

After Tyrone was gone, I had SL leave, so she wouldn't see me picking his body up.

The dead don't look like they do on TV. It turns out that postural muscles are more important to looking "alive" than you might think.

SL will go to sleep tonight, for the first time, with me traveling but without Tyrone at the foot of her bed. She will feel that same silence.

I would give just about anything to spare her that feeling.

But I can't. The first step to getting past an addiction is to admit you are powerless.

I guess I'm addicted to death. I am powerless over it.

At some point, I'll be the one passing on. Perhaps it will be on a motorcycle, with me making a mistake in traffic on the freeway. Or with some distracted soccer mom leaving me a greasy streak down the side of a road, while talking to her nanny on the phone.

Or, maybe it will be cancer. Or some other disease we could have cured if we had not chosen to invade Iraq.

I don't want to die alone.

That said, SL deserves better than to sit next to me, waiting for a nurse to tell her I'm dead. She should be sitting somewhere comfortable, watching the sun go down, with a glass of rose champagne in her hand, when some nameless, faceless functionary comes to tell her I'm dead.

And I really hope that Elisabette doesn't have to hold my hand when I die. I don't want that image in her head.

In the mass production zone

But there's nothing I wanted more than to be with JR, Minnie, and Tyrone when they died. I wanted them to know that I was grateful for their time in my life, and to offer what comfort I could.

So, perhaps I'm selling SL and Elisabette short. Or, perhaps, like Tyrone, I want to be doing my job, to the last, to protect them from what I can, shield them with my body, my heart, my soul.

Protect them from seeing Death. My death.

It's dark outside the plane now. We have outrun the sun. And, perhaps, it's time for me to put my keyboard down.

I've made myself cry. Did I make you cry? Or just make you want to go hold someone so tightly that they will always remember.

Remember the happiness, the love, the fullness. Just, maybe, for a time, feel power over death.

Bring back the sun. Or thrive in the starlight.

'cause your friends don't dance
And if they don't dance,
well, they're no friends of mine
--Men Without Hats

Saturday, August 20, 2011

Trim the plants around the cameras

Well I guess you scared me too
--Concrete Blonde

First, let me shout in jubilation that the West Memphis Three were freed today, 18 years after a huge, terrible, awful miscarriage of justice.

Sadly, instead of walking out of prison as clearly free men, they are walking out of prison under a most bizarre legal arrangement that has them proclaiming their innocence, while the criminal justice sees them as guilty. Kind of.

But, at least they are out of jail.

I wonder what each one thought as they felt the Arkansas sun on their face, smelled the late summer honeysuckle, heard cars go by. All the same as when they were in prison, but, this time, wearing street clothes, with nobody to stop them walking across the street, no guns, no guards.

Do you think they are slightly scared? The world of the last two decades was awful and unfair… but they were used to it.

I wonder. Is the joy balanced with terror?

I started looking for a warning sign

I'm sitting at my desk in the corner of the main room of my company, listening to music. Pretty normal day… well, except that most days I don't get to spend much time at my desk, but let's ignore that for now, ok? You'd normally find me drinking coffee, or water from my water bottle.

// Side note: really, don't use the plastic water bottles, think of the landfill waste!! //

However, at the moment, I'm drinking neither water nor coffee. I'm drinking a Canada Dry ginger ale from a can.

The can is showing ads for characters in the upcoming movie "Captain America".

Am I the only one that perceives a bit of irony in a drink called **Canada** Dry showing ads for Captain **America**? An unusual marketing arrangement.

I'm quiet you know.
You make a first impression
--Howie Day

We are hiring like mad at my startup company, ZestCash. (Yes, if you're interested, we are hiring software engineers, machine learning experts, product managers, and relationship managers.)

This week we interviewed 4 software engineers, two product managers, and two senior non-engineering related people.

Wow. My calendar looks worse even than it did at the end of my days at the Big G in the sky.

But it is a really good use of time. Perhaps the single most important thing a startup CEO can do.

It's super important to hire well at small companies. Hiring matters everywhere, I think, but mediocre employees can bounce along at large companies. Weak hires don't add much, other than cost, but they also don't risk the company's existence.

At a startup, bad hires can sink the ship. A single bad hire sets a cultural tone that excellence doesn't matter. One bad worker makes everyone else work less.

And, at least at my startup, there is always a lot of work to do. If everyone works a little less, we are less likely to hit our targets. Thus, bad hires make it less likely that the company will thrive (or, I guess, even survive).

If you'll let me use a bit of "HR speak", talent is the whole game. I'm obviously not the only person who thinks this. Mark Suster agrees. Dave McClure does as well. Marc Andreessen has a slightly more nuanced view, but doesn't contradict the point.

If your product space is dumb, or is too small make a viable company, then perhaps talent won't be enough. But, otherwise it's all about great employees.

I believe that, given the right talent, you can find a good product-market fit. Thus, I spend a lot of time on hiring. We had the same general view at Google, where I spent at least 25% of my time dealing with hiring in some form or another.

// Side note: It's pretty cool to be able to compare my little startup to Google! //

Over my career, I've hired literally thousands of people. I've fired a few. I've laid off many… thanks to Schwab, mostly. I have a group of folks that tend to move from company to company with me.

I get asked a lot how I think about hiring, what should candidates say to get my attention, or how resumes matter, etc.

Thus, this blog post: How does Douglas think about hiring.

First, a few words from our sponsor. I don't read cover letters. I don't care what they say. Maybe professional recruiters do, but I don't. Don't waste time writing one. Also, I don't care if your resume is really pretty. Fonts, etc., don't impress me. It's amazing what Microsoft Word can do for you. Keep it simple.

// Side note: If you write your resume in LaTeX, I will be impressed. If you don't know what LaTeX is, don't worry about it. //

OK, so what do I care about?

On a trivial note, I care about spelling. "Principal" is not the same as "principle". Make sure you know which is which. I recently saw an applicant who had experience in "discreet math". I don't think I care that you do your computation behind closed doors.

And, by all the saints, spell check the stupid document. Really, Word will put a squiggly red line under a word it doesn't recognize. Look for the squiggle. If Word doesn't recognize the word, you're probably spelling it wrong.

// Side note: The squiggly also might be under some Google-related term. Not sure why that would be… //

For alternative spelling advice, do a Google search for the word, and pay attention to the spelling correction suggested.

Lest there be any confusion, let me be clear: If you can't even QA your own resume, you almost certainly won't be working in my code base.

My company writes virtually all of its code in Ruby. You'd think, therefore, that you need to know Ruby to get hired. Not true. We will let you interview in whatever language you like (as long as one of us speaks that language!).

But you really don't want to put down that you know Ruby if you don't. We are going to ask you to write code on the board. If you say you know Ruby, we are going to ask you to write in it. And if your code stinks, it will reflect badly on you. Stick to a language you know. I have nothing intrinsically against Java, C, C++, or even (horror) C#. You get my point.

If you come interview at ZestCash (and did I mention we are hiring?), here's what will happen. We will read your resume (that you submit via our jobs page).

First, I'm going to look at your education. It's not a determining factor, but it does have signaling value. Did you go to some super snooty school and manage a C average? I'll be interested in what you did instead of studying. Did you go to a really not-great school, but come out with a 4.0? I'm going to be excited to talk to you.

// Side note: By the way, I went to a second-tier undergraduate school. My overall GPA was pretty good, but nothing to write home about. //

Some of you
might still be in that place

I'm going to look at your job history next. Have you jumped from job to job, staying a year at each place? I'm going to wonder what was firing your jets to leave so quickly. The lifespan of a job seems to vary by industry. I once interviewed for a job at a hedge fund (no, don't ask, it was a bad idea from start to finish, but that's another story). The head of the hedge fund asked why I had so many jobs, for such short periods. I have a series of 4 - 5 year stints. Although pretty normal for tech companies, apparently it's abnormal for financial services?

I'm particularly looking for 12 month jobs. I've got one on my resume. Feel free to ask me about it. I'd ask you.

And when you answer, I'd suggest that you not spend a lot of time throwing your old boss under the bus. Every success has a lot of factors… and failures do as well. I'm going to respect you more if you talk about what went wrong including what you didn't do well.

And I'm not going to be blown away if your answer is "well, job #2 gave me a 20% raise." I am not going to get into a bidding war with your next employer -- that's not how we work -- so I'm now wondering if I'm going to be able to keep you more than a year.

Then, I'm going to look at the path of your jobs. I want to know you did interesting things at each job. I don't care if you decided to go into management -- not everyone does. I don't even care, much, about getting progressively bigger titles. I care that you are trying different things, learning at each level, teaching at each level.

Sadly, I may also have to look at your visa status. Startup Visa program anyone?

OK, enough paper stuff. Time to do some work.

One of my engineers (or product managers) is going to call you. This "phone screen" will be about 45 minutes long.

// Side note: If I call you, the call will probably be a bit shorter, since I talk really (really) fast. Don't take it personally. //

The phone screen will cover your experience -- so be prepared to walk through your resume. This is probably the only time you'll have to do so! Prepare!

Then, you'll be asked a technology architecture question, to see if you can think through some tech gunk. We probably won't ask you to write code, since that doesn't work over the phone, but we might ask you to, for example, walk through the architecture for building a crawler.

After that screen, the interviewer will put a score (we currently use 1 - 4 scores, with "1" meaning "no way" and "4" meaning "can the interviewee start tomorrow?"), and a few sentences of feedback into our tracking system.

If the scores are south of 2, we'll probably end the process here.

If the scores are around 3 -- as most are -- we will have a second phone screen with you, with a different person asking similar questions. At this point, we will either invite you to come in for onsite interviews, or we will end the process. There's no option for additional phone screens.

Our onsite process involves interviews and homework. If we are going to bring you in to talk to us, we are going to give you a piece of homework to do in advance. The goal is for you to spend a few hours attacking some problem, usually a mix of architecture and a bit of code.

The first 30 minutes of your onsite visit will be presenting this homework to a group of us. We'll ask some questions, but we don't expect (or want) polished presentations. We want to see some code (that runs, by the way!) and a reasonable architectural diagram. The 30 minutes lets more of us see you think and talk, and gives you the chance to see several of us. We want you to want us, too!

After that, you will have a series of interviews. You will have at least two engineering interviews. Each will be 45 minutes long. At least one will require you to write code on the board, and the other will ask you to solve some architectural tasks. These interviews will be hard. If you aren't sweating a bit at the end… something went wrong.

At some point during the day, you will be interviewed by a non-tech person. That interviewer will be testing for cultural fit. Even if you're brilliant, you may not be the best fit for our culture. And, to be fair, we may not be the best fit for you! To be a Zestian, you have to be comfortable with our culture. We eat lunch together every day -- and nobody wants to sit next to someone boring, right? Everyone talks about everything. Although people have focus areas, they aren't wearing blinders. You might be asked a technical question by a finance person, or be asked to do some (seemingly) random analysis for me. It happens. We aren't a great place to work for people who struggle to get things done when distracted.

At the end of the day, normally, you'll see me. If my calendar doesn't allow it, let me apologize in advance. I'm going to ask a few questions, and give you a chance to ask about the business. You should have some questions. It looks bad if you don't.

Then, you'll be done! W00t! You can head off to the bar!

But we can't.

Right after you leave, we will all gather in a conference room. Each of us will give our feedback -- including that 1 - 4 score. Our goal is to reach consensus on whether you are smart or not. You could have tanked on the algorithms question, but done a good job trying to figure it out. You could have completely whiffed on my thinking question, but been interesting in how you tried. You get the picture -- we cared less about the answers than the ways you tried to answer.

Truly, the journey is the destination.

We will get back to you within 24 hours, in almost all circumstances. We appreciate your time, and want to be respectful. And, if we like you, we want you to like us!

Hopefully, you'll love the offer, say yes, and we will have a great time together solving really hard problems.

(Did I mention we are hiring?)

So, if you've read this far, you have a pretty clear picture of how we hire. But is it clear why we hire that way?

1) We hire for horsepower. I can teach you skills. I can't teach you to be smarter.
2) We want you to be a well-rounded person.
3) I want you to believe in what we do -- we are not just a company, we are a mission.

That's right. Not just another company, a mission. Pay attention now, let's talk about the mission…

Our product? A short-term loan that is about 50% the cost of the alternative.
Our mission? To save the underbanked billions of dollars.
The best part of every day? Talking to a customer who, because of our loan, can pay their rent, fix their car, pay a health care bill, buy school supplies for a child. You get it. We care about our customers.

Do you want to join us?

Rise above
we're gonna
rise above
--Black Flag

Friday, August 05, 2011

Customer Service matters

One more time,

loud as you can


I was in Australia for a week recently, carrying (sadly) all of my devices. I was carrying a MacBook Air, an iPad, and my iPhone. I hate traveling with that many bits of hardware, but, c'est la vie, I guess.

I was carrying the Air in case I needed to project stuff during my talks. Turned out I didn't, so that was a waste, but I didn't know that.

I was carrying the iPad because it's my primary device now -- I use it to read, to do brief emails, listen to music, etc.

I was carrying my phone to have an easy to use phone (yes, I could have used GVoice, FaceTime, etc., on my iPad, but I haven't bothered to set any of them up. Sorry.)

Now, I normally turn international data roaming off on any device I carry, relying on wifi wherever I am to grab content and email. However, this trip included loads of time where there wasn't any easily available wifi.

Ergo, I left data roaming turned on for my iPhone.

I'm grateful that they show me the way

'cause I could never know

--Steve Taylor

So, if you do this, you land in , and immediately get a text telling you that "international data rates apply of $ per MB".

OK, thanks, I get it. Although I appreciate the warning, I deliberately ignored it. I needed my data, so I was resigned to paying a lot for it.

And having ignored the warning, I ignored the data rates themselves. I didn't care about the cost at that moment, I cared about the functionality.

On my 3rd day traveling, I got a text (from some random number) telling me that "[my] international data usage is high. Please call AT&T at ". OK, I've got other things to worry about, but I dial the number, put my phone on speaker, enter the dozens of different numbers asked for by the IVR, and wait.

I'm not just sitting here during this wait, I'm actually trying to work. But every time a voice comes on the line, I stop what I'm doing to pay attention in case it's actually a human.

I wait about 15 minutes for a real human, but have listened to recorded humans about every minute during the process. With such a high rate of interruption, I got virtually no work done.

Heaven forbid

you end up alone

and don't know why

--the fray

At 15 minutes, another recording comes on, but this one tells me what the call is about -- they want to sell me a new data plan. They have wasted my time in order to sell me some new product.

I hate that. I don't care about AT&T's sales goals at this moment.

// Side note: I don't care about AT&T's sales goals at ANY moment, actually, but I was particularly disinterested at that moment. Back to the main line now. //

Having realized that this is a sales call, I hang up. The (annoyed) finger having writ, moves on.


The Neighbor of the Beast

--a T-shirt

A few days later, I get another text from that same random number. This time, the message is "due to your high data usage, we are turning off your data plan. Your data access will not be turned back on, even in the US, unless you contact us at ".

OK, now we've changed from a minor annoyance to a fairly major one. So I dial the number… and have the same experience as before. This time, however, I wait for a real human -- it takes about 20 minutes.

The real human that I talk to is a customer service agent. Now keep in mind, my only goal is to get my data plan turned back on. Nothing more, and nothing less. Not a terribly hard problem.

However, the customer service agent can't turn it back on. I *must* sign up for an international data plan to get the data turned back on.

So, I listen to the sales pitch, and sign up for the cheapest plan. This gets me my data back. Yay!

And, it yields *less* revenue to AT&T. The amount I paid for my plan (even annualized) would be less than the expected international data costs would have been, given my travel plans.

Yes, that's right, not only did AT&T annoy me, they yielded themselves less revenue!

Some marketing person should lose their job, along with whoever does financial planning and analysis at AT&T. Really, the core value proposition is to generally maximize revenue while making your customers happy.

In this case, neither goal was hit. W00t, AT&T!

Know that these things

will never change for us at all

--Snow Patrol

But there is one good aspect to this whole process: The agent I talked to. He recognized immediately that he couldn't do what I wanted (ie., turn my data back on), and gave me a clear explanation of what his options were (ie., sign me up for some data plan).

Although counterintuitive, this is exactly what customer service should be able to do. No "escalate to my manager" to make a decision, no "here is the smoke-and-mirrors to get you to be happy again". The agent had one option in the situation and was clear about it.

Sadly, I didn't pay attention to his name. However, if any of you work for AT&T, go find him and tell him he did a great job. Good customer service is worth its weight in gold.

At ZestCash (my company), we have gone a long way to hire really (really) talented people to provide our borrowers service. We call them "relationship managers" not because we are playing buzzword bingo, but because it captures how we want them to act.

An RM on the phone has the ability to make almost all changes to a borrower's account. There is never a need for an RM to escalate to make a decision, and our RMs know their options for an account before the call even starts (we have great systems, not surprisingly).

Although I'm hugely annoyed at AT&T, my experience with their customer service was great! They ended up with less revenue from me, but that's a marketing problem, not a service one!

I hope -- trust -- that if anything goes wrong, my customers feel that their relationship managers care about them, are exceptionally clear, and make good decisions.

Wednesday, June 22, 2011

Maybe the gypsy lied

All you have to change
Is everything you are

5 years ago today, Jeanne Michelle Russell won her battle with cancer. I wrote about it here.

I was sitting with her, at about 1:30pm pacific, holding her hand, in the little room down the hall.

She asked me if I was ready.

I said yes.

I was an idiot.

And who shall wear the starry crown?
--Alison Krauss

I met Jeanne -- I called her JR -- when I was at Schwab. If you're interested in the story, it's all laid out, in embarrassing detail, in my blog.

From this distance, it all looks different. I feel differently - I've released on some of what I got wrong. I've gotten more upset at myself about other things. I understand some things more, and find others even less clear, fuzzy, as if in a haze from a distance.

It feels a bit like a sharp object wrapped in cotton for shipping. If you hold it "right", it doesn't hurt.

But most definitely if you grab it by the point, it still draws blood.

What made you forget
that I was raw?
--LL Cool J

Dear Jeanne:

All these years later, I still celebrate your birthday. I remember how your eyes shone when I brought you flowers, or bought you a present, or said you were lovely.

And I still mourn the day you died.

I know that Minnie is with you now. I hope you are remembering that she likes to rip the squeaker out of her toys, and vivisect the little stuffed ones.

Get her lots of toys, she deserves them.

And don't forget to keep her away from the hot dog buns. Remember the year that she grabbed the bread off of the little table that we put beside the bench in the back? How sick she was for days? It was sad… but, really, it was funny. You know it was, even though you didn't say so at the time.

When you lose something
You can't replace.
Could it be worse?

Brown is lying at the foot of the bed, panting. He's getting lots of love. Lots of treats -- he still likes the Greenies best. Lots of pats -- but he still hates to be groomed. He eats mostly soft food these days. I know, I'm spoiling him. But he's a good boy. You'd be amazed at how much more pack-like he's become. He wants to hang with us, he kisses us, he gets bummed when I travel.

He's become a bit like Minnie, actually. The best of him, with a touch of her. It's wonderful.

I think I'm better too. Perhaps a bit more like the man you always thought I was. I'm a little calmer, and a little more understanding. I hug more often, and tell people that I love them more.

I learned that from you.

You'd really like Sonya. You both think the same things are funny, and both love wedding magazines. Really, what is that all about? Regardless, you'd be pals.

And Elisabette is remarkable. Luckily for her, she doesn't look or act like me at all. But I'm still going to read to her, kiss her, and try to do my best with her. I wish you could say hi to her.

So much of what's good in me came from you. When I'm sad, I try to think about all the good you did, and all the people you touched.

But, really, I just feel my way into the parts of me that wouldn't exist without you.

I shall never forget you. Five years on, I'm still crying.

And suddenly
I become part of your past
--The Fray

Thursday, March 03, 2011

Musings on technology, product, and customer service

Dressed up like a million-dollar trouper
Trying hard to look like Gary Cooper

There's an article on Slashdot today talking about equity trades "in picoseconds". Fascinating stuff. Trading systems are complex -- there may, or may not, be a central arbitrator of trades. Usually there are only two parties to a trade, although there can be more. (I don't think there can be less, under most assumptions of mental health.) There must be a way to settle trades between parties.

For a trade to matter, two parties have to agree on a security, a price, and a clearing method. Clearing methods are usually automatically selected, so that shouldn't take any extra time.

But you still have to find a security and a counterparty.

Trading at picosecond speeds? it takes about 3 picoseconds for light to travel a millimeter. The average ethernet cord -- the thing that goes from your computer to the wall plug -- is 5 feet long, which is about 1,500 millimeters. So, it will take about 5,000 picoseconds for a single pulse of light to get from your computer to your wall. Add on a substantial amount of time -- in milliseconds, anyway -- to get through your router and to your ISP.

And let's not forget about the time it will take your packet to traverse them-there-Interwebs to the arbiter (or to the counterparty). I'm sitting on a good network, and I'm about 90 milliseconds from They have pretty good connectivity. Like, really good. And they are 90 milliseconds away.

90 milliseconds is about 90,000,000,000 picoseconds. (Just to put that number in perspective, there are about 197 million square miles on the surface of the earth. Which is about 500 times fewer than 90-milliseconds-as-picoseconds.).

We may be talking about trading in picoseconds, but, if we are, we aren't making any sense. In other words, where's the Tardis?

Your faith was strong
but you needed proof
--Leonard Cohen

I'm crazy into bags -- laptop bags, messenger bags, suitcases, whatever, I like them. And I own an unreasonable number of them.

But, still, I'm always in search of the perfect .

Lately, I've been wanting a way to hold my phone onto my shoulder bag strap, so that it's easier for me to talk on my headset when in an airport. I use a binaural wired headset, like the one that comes with the iPhone (but not that one, I use one with more comfortable earplugs and a longer cord).

Even though my headset has a longer cord, it still isn't really long enough for my phone to be in my pocket while using it.

Hence, my search for some way to hold the phone on/in my bag. The obvious solution is a sock-like-thing that attaches to the strap of the bag.

Obvious, yes. Easy to acquire, no. I've tried several. Each has some failing -- too big, so the phone falls out; bit that holds it into the strap too weak, so it falls off. I've almost given up on the whole idea.

Until I found Waterfield bags ( They sell a few things that I really want, including a phone holder. (They also sell a cool gear bag, which I also use, but that's not the point of this post-let).

I ordered the phone holder last week, in black. Because black is slimming, you know.


I find it hard to tell you,
I find it hard to take
--Gary Jules

A day or so after my order, I got an email from Waterfield's customer service, letting me know it wasn't available in black, and apologizing. However, they could ship it out to me immediately in a different color.

I didn't care about the color that much, so I picked one that was available.

And about 5 minutes later, I had another nice email from customer service, thanking me and apologizing again.

I got the sock-thing the next day.

I am blown away by how friendly and competent the support folks were. The response was fast, accurate, and left me feeling good about them not having the color I wanted.

Well done, gang; few places deliver so well.

I really (really) want to support places that get customer service right -- so please open a new tab in your browser, head over to sfbags, and buy something.

And, yes, the phone stays in the holder, and the holder stays on my bag strap. Thanks for asking.

I'll flag
I'll fall,
I'll falter

I hate technology that is used in search of a problem. For example, I don't understand the towel holders in airport bathrooms that use the infrared sensor to know that your hands are in front of it so it can spit out some towels.

First, it never gives enough towels, so you sit there waving your hands in front of the sensor like concert goers at some hair metal band show holding a lighter.

Second, the sensor never works, so you spend a while trying to figure out whether to wave your hands, or leave your hands, or dance a jig in order to get your towels.

How about a simple design where you pull the towel down from the bottom and it rolls out from the holder. It can be perforated to ensure you don't get enough towels if that's an important design goal, but at least one knows what to do with it.

The technology doesn't help, why put it in place at all?

And don't get me started on full-body scanners in airports.

Technology is great for some things, but not for everything. We should focus the tech-bits where they can help, make our lives easier, not where we can just think of putting something.

I wonder how many picoseconds it takes to spit out a towel.

Science fiction,
double feature
Dr X will build a creature
--Rocky Horror Picture Show

Saturday, October 16, 2010

The menu from the restaurant at the end of the universe

Hi to all…

Since ZestCash launched, there's been a lot of discussion about our product, and the space as a whole. We thought it would be helpful to drill down on a few areas of our business, especially pricing and costs.

Come writers and critics

Who prophesize with your pen

--Bob Dylan

On our site, we lay out several options for short-term credit. Readers will note a few key points on that page. Most importantly, credit (of any sort) is more expensive than using savings to buy something.

Even so, people still need credit. A borrower who does not have access to traditional credit could go to a payday loan store or to a pawn shop. According to the FDIC survey of the un- and underbanked in the US, 30 million Americans have taken out a payday loan or a pawn loan in the past year. Clearly there is a need for this kind of credit.

Given that, I believe that people deserve to have credit that is reasonably priced, transparently offered, and fair. I don't believe anyone disagrees on this point.

However, thoughtful people can -- and should -- disagree on exactly how to provide that credit.

There is evidence that payday loans yield a marked increase in personal bankruptcies. There is also evidence that *lack* of access to payday loans leads to higher bankruptcy rates and higher costs for alternative credit-related products like bounced checks.

Both articles point to the need for credit, and the fact that people who get payday loans are right on the edge of financial ruin. Payday loans could push them over the edge -- or not having them when needed could push them over the edge.

The data is unclear. So, let me start with the hardest, most important question: What is a fair interest rate? This is hard because the question has so many facets, and so much importance to social policy.

You'll be singing in the rain

Said don't stop to the punk rock


First of all, let's be clear: APR is a good measure to use to compare credit options. It's been the standard for 30 years, and has been used to drive public policy discussions across many credit domains.

On our site, though, we suggest that total loan cost can also be a useful measure for short term credit. Let me explain why we say so.

For a given loan, with the same interest costs, APR varies with the length of the loan. Generally, all other things being equal, a shorter loan with the same out of pocket cost for a borrower, has a higher APR. If you were to stretch exactly that same loan out over a longer period of time, it would have a lower APR.

However, these two loans would have exactly the same out-of-pocket costs to the borrower, despite different APRs. The borrower will have paid the same dollar amount of interest for both loans.

Mullainathan and Shafir (2009) point out, in fact, that APR may not capture the way that payday loan borrowers actually make credit decisions.

This doesn't make APR a bad measure. It means that, under certain circumstances, looking at the cost -- the out of pocket costs -- is important to consider as well.

This is why we suggest looking at total loan cost on our site. We aren't trying to hide APR -- in fact, we list it on each page of the application -- we are just trying to help our customers understand their actual out-of-pocket costs.

There have been several attempts to set reasonable prices for credit. Many advocacy groups and governmental organizations have suggested that 36% APR is the appropriate interest rate for small dollar, subprime credit. This rate has been embodied in various state and federal regulations over the past few years.

The FDIC did a two-year study, the Small Dollar Pilot Program, to examine bank and credit union alternatives to payday loans and other high cost loans. They asked 28 banks to give loans to consumers at 36% APR or lower. The loans ranged from under $1,000 to $2,500. Overall, more than 30,000 loans were made, with loan durations of 12 - 18 months, and total loan principal of $40m.

The pilot showed reasonably low default rates, which is promising, and showed that traditional financial institutions can, with appropriate incentives, offer less expensive credit. This is very promising for the credit industry.

However, it's important to recognize the limitations of the study as well. The overall pilot -- the total loan principal -- was relatively small. This makes it harder to predict how the products would scale into the broader market.

Also, the banks didn't make profit on the loans: "However, given the small size of [loans], the interest and fees generated are not always sufficient to achieve robust short-term profitability." [FDIC final report, page 5].

Banks currently offer higher cost short-term credit products, especially overdraft protection. There is evidence that banks and other financial institutions don't offer lower cost short-term credit products because they compete with overdraft-related products.

Given that the FDIC pilot loans weren't profitable and compete with banks' other revenue streams, it's not surprising that the pilot participants aren't really marketing the products any longer.

This doesn't mean that 36% is the wrong rate. It means that, with current underwriting, it is difficult to make a profitable product at 36% APR. Better underwriting yields reduced default rates, which, in turn, should yield lower interest. This suggests that we should advance the state of the art in underwriting.

There is hope that underwriting can lower the costs of borrowing. BillFloat shows us that hope. BillFloat offers short-term loans for a low interest rate -- they are normally under 36% -- to pay a monthly utility bill. They are able to charge this little both because they are exceptionally smart and because loans to pay off a month of an ongoing utility bill have very low default rates. You want to keep your utilities on and you're accustomed to paying the bill, so defaults will be rare.

Although this isn't our market -- we offer more general cash loans, which is a segment of the market that has dramatically higher default rates -- this makes us optimistic, knowing that there is at least one company that is doing it.

The kids are all wrong

that's all I ever really learned


This is why we are working so hard on underwriting. If we can underwrite loans more effectively, we can offer credit at lower than current rates. Underwriting for a general loan -- without a specific purpose and with less historical data -- is harder; these loans have a higher default rate, and hence require higher prices. But if we can underwrite better, and lower the default rate, we can provide credit at lower costs and build a good business.

This is important: I am trying to build a business. I am not trying to build a nonprofit. Nonprofits can have great impact on the world, and many social changes have begun with nonprofits. However, we believe that long-lasting change in the financial services world requires a sustainable business model, one that generates profit.

Building a profitable business means pricing goods and services above their cost of production. In this space, the cost of production is the sum of how much it costs to acquire customers, the cost of the money lent out, and the rate at which people fail to pay off their loans. This last factor -- default rate -- is key. As default rates increase, pricing must increase as well to cover those costs.

This is what yields such high prices on subprime loans. We price our loans to cover defaults, based on our underwriting. We are already markedly less expensive than payday loans, and our aim is to be even less expensive over time.

Market pricing is very important; pricing competition is good. In fact, the Kansas City Federal Reserve found that competition between payday loan providers yielded lower prices. Adding our alternative products into the mix should increase downward pressure on payday loan prices, yielding substantial benefit to borrowers.

There has been a bit of confusion in some of the comparisons between our products and payday loans. This makes sense since there are a lot of moving parts in these products.

Before going into the comparison, I want to be clear that we understand our loans are expensive. There are cheaper options, and we explicitly recommend to our customers that they should use less expensive options if they can. To fail to make this recommendation would be neither transparent nor fair.

I was just passing through

must have been seven months or more

--Creedence Clearwater Revival

If a borrower doesn't have other options and needs a short-term loan, there are only a few options.

If a person wants to take out a payday loan, and will pay it back within two weeks -- without rolling the loan over -- then that person should not take out a loan from us. This is why we don't offer loans for less than two months -- our loans will not be less expensive for shorter loans than that. Again, we call this out clearly on the site.

Our loans are normally less expensive than other options, like pawn shop loans or payday loans. How can this be true? It's true because people don't normally pay their payday loans back after only one cycle, they normally require several cycles to pay them back.

If you don't have the money to pay your loan back after the first cycle, the payday loan provider will lend the money -- the same money -- to you again, but charge you fees again.

According to the Center for Responsible Lending, people roll over their loans 9 times. Other researchers have suggested that most people roll over their loans 6 times on average. Like it or not, payday loan borrowers normally roll over their loans.

Each time the loan is rolled over, new fees apply, but the principal remains unchanged. So, if you borrow $400 and pay $80 in fees for the first loan, two weeks later, you will pay another $80 in fees, but still owe $400 in principal. And so on until you pay off the entire principal.

If you roll this loan over once, you will pay $160 in fees ($80 for the loan, and another $80 for the rollover) and then pay an additional $400 (the principal) at the end of the second period. Thus, in one month, you will have paid $160 in fees and $400 in principal.

If you roll the loan over 4 times, you will have the loan for 8 weeks (4 loans at two weeks each), and will pay $320 ($80 * 4) in fees. At the end of the 8 weeks, you will also pay the $400 in principal back, for a total of $720.

Ok, to use round numbers, let's assume you roll your loan over 7 times. That will get you a total loan duration of 4 months (two weeks for the first loan, and 7 * 2 weeks for the rollovers, for a total of 16 weeks). You will owe $400 in principal -- at the end. You'll pay $80 for the first loan, and for each of the 7 rollovers, for a total of $640 in fees. All in, you will pay $1,040, since you pay both fees and principal back.

And a ZestCash loan? If you borrow $400 for 4 months, you will pay $80 every two weeks for the life of the loan. Just like a payday loan, right?

No, wrong.

Because at the end of the 4 months, in a ZestCash loan, you have paid off BOTH interest and principal. So, you will pay ZestCash a TOTAL of $640 (8 * $80), which includes the principal. Because ZestCash loans are installment loans, you pay principal and interest off at the same time.

So, in this example, our loan is $400 less expensive ($1,040 - $640) than a payday loan of the same duration.

The photograph on the dashboard

taken years ago


So, that's our story. People need credit, and sometimes their only option is a relatively expensive short-term loan. Although a 36% APR sounds great, it has proven difficult to create general purpose subprime loan products at that interest rate. There has been relatively little market pricing pressure on more expensive loans, so prices of the higher cost loans are not dropping.

I got involved in this space because I think this is a really important problem to be solved. I believe that the current market is not putting sufficient price pressure on subprime loan products, and that adding a markedly lower cost product that is sustainable will push prices down across the entire space. This cost decrease could put literally billions of dollars back into the pockets of subprime borrowers, who don't have a lot of extra cash to play with.

This is a complicated and controversial business to be in. However, if nobody tries to find a solution, nothing will ever change.

We are trying to find a solution.

We think entering the market at a significantly lower price point, and directing our resources toward developing better underwriting models, is the best way we can do that.

We're thrilled people are talking about the problem -- more ideas always yield better outcomes.