Tip:
Highlight text to annotate it
X
OLIVER FRANCIS KOCH: First of all, thank you very much for
inviting us and being interested in what we have to
discuss here.
A little background information.
I, myself, am not from Iraq, even if it
might look like that.
But I work for the German Development Corporation as an
external consultant.
And we basically started a project in Iraq which was
supposed to help develop the Iraqi private sector, which is
very weak after the war.
There are no private businesses, really, so we work
in different segments, like automotive, IT, and so on.
And my colleague Mohammed Al-Samarraie and I, we work
specifically in IT.
So our job was to do an assessment to see where's Iraq
right now, regarding IT?
Where can it go?
Is there potential?
Is there interest?
And at the end, basically--
our conclusion was after three months, OK, the sector's still
quite weak.
It's still quite young, but there's a lot of potential.
So what we decided on together with the German government to
set a program which focuses on facilitating entrepreneurship.
That's where I met Abraham, for example, with initiatives
such as the Startup Weekend or also an innovation
competition, which we conducted.
So when we actually started this innovation competition,
one of the major issues that we saw is people or students
who were just lacking this mindset of going out there,
doing your own business, starting your own projects,
and just pushing your ideas forward.
So we actually decided on getting a couple of trainers
from Germany that would show them a little bit, OK, what
state of the art methods are there from Silicon Valley or
so, that you could use to facilitate creativity, to
capture your ideas, and basically to
turn this into business.
And one of the biggest issues there, when we trained them,
was that we talked about Facebook.
We obviously talked a lot about Google, too, but people
couldn't really relate to that.
They're like, OK, this is Iraq, man.
Really, this is too far away.
So we actually thought about it.
What could we do to ease this transition?
And that's when we sat together, Ralph Magnus from
Berlin, Christoph, Mohammed, and myself, and two others.
And thought, OK.
There's actually a really large demand
growing in food delivery.
Meaning a lot of people or cultures developing towards
ordering a lot of food from outside.
But the main issue is right now the way it works is
someone would call, OK, listen I live between that hotel,
this pharmacy--
and that's how it works.
And they keep calling you, and it takes you on average an
hour or two to get your food, and it's cold.
So that's a real big pain point for the people there.
And at the same time, I think Abraham already said it, we
operate in the north of Iraq, which is also considered the
Kurdistan region of Iraq.
And it's a little different from the south.
Currently, it's a lot more stable, the security situation
is a lot more stable also, and better.
So things such as internet penetration up there are a lot
higher than they are in the south.
So we're talking about 40% in Kurdistan.
Kurdistan, Iraq.
Just some background information.
Iraq has about 33 million inhabitants.
In the Kurdish area, we have about 5.5.
So this brings me more or less to our elevator pitch.
Easy Bites is more or less an internet startup which focuses
on providing location-based delivery services
to the local community.
At first, now we're focusing on the Kurdish part of Iraq
simply because that's our test market.
It's secure, it's more stable, I, myself, am allowed to
travel everywhere, and it just makes things a lot easier.
The mid to long term objective is, obviously, to scale this
business model, take it to Baghdad.
I mean, Baghdad has a large potential, itself.
It has--
sorry, Mohammed--
7 million inhabitants.
And a lot of the problems that we're actually solving here,
meaning not finding addresses, or not being able to choose
your food, and so on, have a totally
different meaning in Baghdad.
Baghdad, obviously, you see the news.
There are a lot of explosions every day.
So people were telling us, hey, you know what?
For Baghdad, this has a totally different meaning.
This actually means I can stay at home, and I don't have to
expose myself to the threat of bombs, and I would get the
food to my house.
So we're talking about a totally different value
proposition that we're working on there.
So we don't really want to engage into an hour
presentation and bore all of you, so our idea is to
highlight some business aspects of how we operate,
what our idea was, what we've gone through, then go a little
bit into the techie part.
Ralph will cover that.
And after that, just scoop into a discussion, if that's
OK with all of you.
So business aspects.
Most of you probably know Eric Ries' work, the Lean Startup.
Maybe you've heard of it, maybe you've not heard of it.
It basically, if you want to summarize it, for us it has a
fundamental meaning.
Because the idea is basically just that you say, OK,
startups are totally different from established enterprises
or companies.
Because you don't know the market, you don't know if the
product's going to work, and so on.
So basically, all you have is a conglomerate or a long list
of assumptions.
And your goal should not be to make money right away.
Well, it should be.
But you should actually focus on
generating a lot of learnings.
You want to try to turn all of these assumptions that you
have into learnings.
Because only then, OK, now I can scale my business model.
Now I can invest all my millions or billions into
marketing, and it would actually have an effect.
So the principles that we were guided by most of all, I would
say is this thing that Eric Ries proposes.
It's like a cycle, a scientific cycle.
Meaning you have an idea--
let's say for us, for example, it was a
food delivery service.
So you would build something that people could actually
understand the idea.
So for us, that would be--
we call it a minimum viable product.
Meaning it could be a video, it could be a cartoon, where
somebody just gets the business idea.
What exactly is this about?
And after that, you try to measure, OK, and how far is
this interesting to people?
So what we did, we basically placed this digital MVP in a
video on a website and told people, hey, you know what?
If you're interested, pre-register for a free beta.
And it gave us the feedback, is this actually something the
local community wants or not?
And after that, obviously, going back to this
Build-Measure-Learn, you would you then say, OK, this
actually makes sense.
And then you start building around it.
And then you go into the next phase.
You actually have a tangible product.
Meaning you try everything on a smaller scale, and then you
scale it after that.
Let's say from a development standpoint of view, or what
Ralph will also be emphasizing later, we really try to focus
on small batches.
We didn't have this large, super master plan which we
started implementing and at the end everything was great.
It was actually more like we had an idea and we try to keep
it really, timewise, really small iterations.
Started delivering very small components first.
I think Ralph will give you some insights.
For example, we don't have addresses in Iraq, so how do
we work around that?
So what we did is, we did a tagging competition.
We basically said, hey, first winner is going to get a
Galaxy, or so.
So people were all hyped.
And we said, all you got do is, we
programmed this little app.
Just go to hotels, to restaurants, and so on.
It checks your location via Google Maps, and you just
create tags.
Meaning, for example, hotel or whatever.
The benefit for us is we are aggregating a lot of data
which is currently not there.
And this data can then be used subsequently in delivery
processes to see, OK, the delivery guy doesn't have to
call the customer and be an inconvenience.
He actually sees straightaway, OK, this is
where the guy lives.
Makes it a lot easier.
Basically, most of the principles you see here are
agile guidelines, like deliver and ship early and
continuously.
One very important aspect, I would say, is also simplicity.
Iraq, or developing countries in general, just have really
large challenges that we wouldn't find here, for
example, physical addresses.
For example, also dealing with people that are just--
this is actually a very good example--
most Iraqis that I've met, they know how to use Facebook,
and they know how to use Google, but
just to find Facebook.
So meaning, they basically use the Google search engine bar
to type in Facebook, and then they click it.
If you ask them to use it in any other context, well, it's
overwhelming.
So talking about apps, all they know is the Facebook app.
So in the start, we tried to come up with our really
intuitive own app, and we realized this
is not working out.
So at this stage, you'll probably recognize a lot of
similarities between our user interface design and
Facebook's.
Just simply because we've got to work with something they
know and they feel comfortable with and they
want to engage in.
That's why simplicity.
Even if a lot of other things that we've been working on,
for us, were a lot more intuitive.
But there's that.
I'm going to skip to a little bit about this
product market fit.
Now, one of the Lean Startup fundamentals is also that you
can use business plans at some point, but it basically tries
to capture everything about your business on one sheet.
This is what you'll see later.
It's very small, so I think I'll have to read out a lot.
The fundamental part of this is that the Lean Startup
method or model basically suggests that, OK, you have an
idea, and this idea, this value proposition, fits a
certain customer segment.
For example, maybe the pinpoints or the pain that
some Iraqis have regarding delivery services might not be
the same as people have here in Europe.
So you're trying to achieve this thing called product
market fit.
It's always building something, getting the
feedback from the users, and adapting.
And this is a cycle which is ongoing.
And something that we just try to
highlight here in a timeline.
For example, at the beginning I explained we did this MVP,
this digital cartoon that we put
online to get some feedback.
Another part was the tagging campaign, where we basically
said, OK, are people actually interested in the product
itself, and how far are they willing and able to engage
with such an app?
Another point was--
what we realized, with restaurants, especially, that
it doesn't make sense just for them to define a delivery
radius around their location.
If we take a look at the city of Erbil or Baghdad or
Sulaymaniyah, there are just places that are a lot easier
to access than others.
That's, for example, hard infrastructure that impedes
making things a lot more easy and flexible for you.
So what we had to come up with at that point, or at some
point, was then, OK, people don't just define a delivery
radius around their restaurant, but they can
actually define different points around the city which
are a lot more accessible.
Which, obviously for them, also means that they now can
define, OK, when is something economically feasible or
profitable for me, and when not?
Because if you were to just say, I'm going to deliver
around the whole city, well, your delivery guy might be
gone for two days.
Just guiding around, going through this timeline,
obviously other issues that we had, and I talked about this,
is that we had an interface at the beginning UI, which was,
for most Europeans-- because we tested with Europeans a lot
at the beginning--
was totally intuitive.
But the Iraqis, we measured through Google Analytics and
so on where people would actually bounce off.
And we realized, OK, they actually bounce off on the
first stage of the order part.
So we redesigned everything, aligned it towards Facebook,
and that just really blew our mind.
Because after that, we had really a bounce off rate of
maybe 10%, and it was about 90 before.
So we're turning the whole story around.
What we're working with now--
this is, I think, very interesting for you to know.
We only started in December, and it is March now.
We launched the actual app in February and the tagging
campaign in January.
So we're very young.
We're still learning a lot.
And issues such as the absence of physical addresses are
still continuously impediments for us to actually improve the
service, to make it better, to make it more feasible for the
restaurants to deliver, and so on.
But for example, our first stage was to work with these
reference locations, meaning the hotels.
So the delivery guy sees on the map, OK, this is
where the hotel is.
This is where our customer is.
In the second stage--
most of you probably know TripAdvisor, where you have
this compass.
And it shows you the direction, OK, 400 meters, and
shows you the direction where you have to go.
We're working on an implementation of that as
second step.
And in a third step, what we're actually doing--
with the local government, about 15 years back, since
then every house was given a unique code.
And this code basically had the first three digits, which
were more or less a sector.
Then there are another-- sorry, Mohammed--
two digits for the street, and the rest is
for the house number.
The funny part is, nobody's really using that.
I can't explain why.
The only thing I've heard that 10 years ago, people actually
got their magazines that way.
If you look at the postal service today, well, it's just
one building, and you go there once in a while to see if
there's a package or something, which is highly
inefficient, obviously.
So we are still definitely learning a lot in this regard.
And I think in that direction with this code in the future
we'll be able to achieve a lot.
I'm just going to briefly explain the business model.
Now, obviously, what we do is we go out there and we talk to
restaurants.
We talk to restaurants, typically those that would
have already established as a delivery service right now,
because that's just the easiest.
You don't want to convince them.
It might take some time.
And one of the biggest issues that they obviously have--
because in Iraq, aside from restaurants, you don't really
have a lot.
There are no nightclubs.
Bars--
OK, there are some bars, but it's just a food culture.
So it's all about food.
So a lot of restaurants are obviously reach their capacity
limit, meaning they have a certain amount of seats, and
they're just booked.
So they actually see the delivery service as an
opportunity of scaling their business.
Of growing their business without having to make a
substantial investment into new physical infrastructure.
And that's where we basically come into play, and we
convince them that, OK, listen, this is something that
you could use.
Test it for free.
So at this stage, we are operating an [INAUDIBLE]
and we have around 70 restaurants in each city.
But we're growing there.
For the users, what we are basically offering is a
location-based service.
Very obvious.
They turn on their smartphone, they see which restaurant is
in their delivery radius, they browse through the menus,
place their orders, and get their food.
I think it's very straightforward.
What might be interesting for you is the revenue model.
How exactly are we making money on this?
There, too, we had to learn a lot.
In Europe, obviously, commission is
something very common.
First time I walked into a restaurant and told the guy,
hey, listen, let's work this with commission.
He's like, are you crazy?
You don't ask that here.
So he was totally offended.
So after that, I was like, OK, what's your proposal?
He's like, OK, we can agree on a fixed fee per order.
So right now we're charging $1 per order, and everyone's
happy and everyone's--
it's just a cultural difference.
Basically, localizing the business to what exactly is
suitable to the region.
I can actually show you this.
On average, the large restaurants have a 150 to 200
orders a day.
So we would say, right now we have in each city we have
about one large restaurant.
Then we have about three medium-sized restaurants, and
the rest is small.
Medium is anything between 50 and 100 orders.
Small is anything between zero and 50.
Zero wouldn't make too much sense, obviously.
But if you think about it, it's basically something--
it's not going to be anything that in this Kurdistan region
you would be making a lot of money per order.
It's actually that it's a volume-based business, to be
totally honest.
And if you aggregate it up right now, we are already--
sorry, got to calculate--
we're maybe at per day you would get--
with the restaurants we have, you could probably achieve
around 2,000 maybe 3,000 orders.
And we're only covering 14 restaurants.
It's a question of scaling that.
There are a lot more restaurants.
The question is also taking a look at large cities as, for
example, Baghdad.
I think the potential is there, but it's a
volume-based business.
You're not going to make 10, 20, $100 on every transaction.
Right now we're covering Erbil and Sulaymaniyah.
We are planning, or we had people approach us from
Baghdad, especially.
From Basra, too, and some other cities like Mosul and
Kirkuk and such.
So generally, we calculated the total available market,
which obviously means, this is the whole pie just in
Kurdistan right now.
Which we estimate annually to be about 7.8 million a year.
The way we calculated, we'll probably try to send around
the slides, if that's possible through Abraham.
The biggest issue in Iraq is there's
no real data, meaning--
I'll give you an example.
A week ago, I was trying to find out, OK, what exactly is
the internet penetration in Kurdistan?
And what you've got to know, there's really a monopoly
regarding the provision of internet services.
So for example, in Kurdistan there's one company called
Newroz Telecom.
And the CEO made a really, really, really
contradictious statement.
It was something like, so in entire Iraq we have an
internet penetration, I think, of 5%, but that
is only due to Kurdistan.
And Kurdistan has only internet penetration of 8%,
and we here in Baghdad have 2%.
Basically, the numbers don't add up at all.
Because I basically reverse engineered the numbers, and I
came up with, well, based on what he was saying, you at
least need an internet penetration of 50% in
Kurdistan based on the population.
So that's what we're working with, right?
We know one thing for a fact, and that's very interesting
for our business, especially, is that the government
actually measures its internet penetration right now by the
number of Facebook registrations in the country.
Which is totally--
it's funny.
At the same time, it's creative, because they have no
other method of doing that.
Especially because the large corporations, they're not
really publicizing a lot of data because of tax
reasons, and so on.
I think you get the picture.
So the question there is, we still have a really large
growing market regarding the internet penetration.
What we also have-- and this is comparing it to Germany, a
real plus for us--
take a look at the demographics.
Nearly 70% percent of the Iraqi population is between
zero and 34, which is a lot.
Which is really a lot.
Meaning we have a lot of potential, actually, that is
going to be coming up here in the next couple of years.
If you combine that with the internet penetration, which is
also growing like crazy, I think our potential target
market is growing in total.
So I would like to hand this over to Ralph, who will give
you some technical insights and how we try to involve some
locals, too.
RALPH BENJAMIN MAGNUS: Yes, thank you very much, Oliver.
Is this on?
Can you hear me?
OK.
So I think Oliver has said so much, so there's nothing left.
Any questions?
No?
OK.
Just joking.
I want just to come back to this slide.
We always have to have a mind that we're talking about a
delivery time here of about two months from the first idea
when I met Oli first in Erbil.
Then we go to implement the first prototype only for the
tagging campaign, and then we extend it in each increment
application.
So for us it was a good experience, this project, to
test our agile approaches.
So we really implemented only these things, only the
features that are necessary for each increment of our
application.
So the change that is always there in software engineering
project comes from serial sources here in our project.
The first thing is that we always had running serial
instances of our application.
We had the demo instance for testing all new features, we
had the staging system for our partners to get trained to
work with our system, then we have the life system.
And we see on the left, we do all the automatic
testing, and so on.
And we get the necessary feedback from our test suite
what we have to adapt to have a functional system to have
all the features implemented in the right way.
Then we have big feedback from our observations.
So, from our local staff in Erbil and Sulaymaniyah that
are face-to-face with our partners, the restaurants,
that installed the system and explained the
system, how to use it.
And this gives us a very valuable feedback for us.
It was, I think, one of the most important channels here.
And the sell channel was the social channels.
So mainly it's Facebook, where we had all our campaigns and
get the user feedback.
And all these we had to handle on the level of our
application.
So we had to react very fast to involve also the
people at the place.
So if we get feedback, so they saw that the feedback is taken
seriously, that we adapt the features, their
ideas, their commands.
So in general, we were able to adapt and implement features
and interests, and so on, within one day
to maximum 14 days.
14 days was never--
I would say one day to one week, we were able to
implement all changes that happened in these
2 and a half months.
Just some examples that those are not only small changes and
small learnings that we have done in this time.
For example, in the beginning, there was a big issue with
some kind of operating system in a very special version that
brings for us a huge problem.
We saw one of our test suites and test environments made
this visible that in this version, if you created a full
screen home link of this application, that your
location is not working.
It's only triggered once, and if you start the app, and then
it's not working anymore.
So thank god we figured out this in the beginning, before
the users started to use our application.
Because if they want to go around and to tag and to tag,
and for us, these tags are very useful, and we
want to have them.
They have a huge value for us.
It would have been disaster if they wouldn't be able to tag.
Because we have metered that nearly 90% are using iPhone.
And this time we said, version of the operating system.
So we were able to handle this in a way that we have some
operating system detection that detects a version that
allows to create a home link or not.
Second thing that comes from the
observation from our partners--
Oliver already has mentioned this--
was that in the beginning we had the prototype developed
like each restaurant can define a specific radius for
the delivery.
But we very fast saw that this is not sufficient for all
partners, so they need the possibility to create several
independent service zones with different charging, and so on.
So this was a very important observation from our partners
that we also could implement very fast.
And observation that comes from social channels, as well
as from Google Analytics where we observe the flow of the
visitors, was that we simply had to simplify the ordering
and delivery process as much as possible, so that it's
really straightforward.
Just three clicks to get your food, and not options here and
add pepperoni here, and so on.
Just a basic menu and left out everything.
So we had implemented everything.
We would be able to do all this, but we
had to leave it out.
Maybe later, next year, or whenever the users can adapt.
So all that I've said is also reflected in the way we
deployed our application.
We had serial branches and instances running.
On the left we have development branch where
Christoph and me are putting all our new ideas.
And whatever seems to make sense goes into a demo server,
where the local staff can test it and can give us feedback.
And in the same time, we had the localization branch, where
we had the local Easy Bite stuff that is directly
involved in translating everything in time, so that
each new feature, before it comes to the live system, is
completely localized.
We included this localization directly in the application,
so it could be done in app.
And until the version 0.2 we had the staging server, where
we trained all the restaurants before they're going live.
So we can ensure that they are able to handle the system,
that they know what to do if there's a delivery [INAUDIBLE]
From the current version there is training models integrated
and the live server where we can switch a restaurant from
training models to live models, so no need anymore to
migrate restaurants from staging to live, just
switching it.
Here you can see wherever this triangle there was automatic
test execution of our test suite.
So whenever we committed to a feature, it has
to pass a test suite.
So beside that we have this very short period of time, we
always took care about quality and tried to follow
test-driven development here.
That's for the deployment.
Christoph will give some additional information
regarding the quality assurance and the architecture
that we have chosen to implement the system.
CHRISTOPH HERBST: Thank you very much.
Hello, everybody.
I'm Christoph Herbst, and I'm also glad to talk to you in
the Google main headquarter.
Can you hear me like this?
Or do I need this one?
OK, hello everybody.
So just I will do very quick.
So Oliver already told you that we try to develop the
application very fast, and we have quickly changing
requirements.
And that's why we need to [INAUDIBLE] testing.
And we see a small screen shots of our testing coverage.
And from the beginning on, we tried to have a good coverage,
of course, but due to the time constraints and to very
limited manpower we have, we are not yet at 100% in each
controller, but we are trying to assume that.
So that's a screen shot from PHP units.
Then we have a mobile application.
And I just want to say a few words about the special
aspects we have to cover there.
This is a simplified overview of our application
architecture.
What you see is that we use a Symphony back end on the
server, which delivers mainly HTML files for cover HTML.
But we can also deliver JSON data in certain circumstances
where it's very important that we have small data that we can
have a good performance.
Then we use a lot of JavaScript libraries.
So standard libraries that you might already know,
JavaScript.
And to have a consistent mobile look and
feel on every platform.
We use jQuery mobile.
And of course, we use Google Maps a lot, and a few others.
So then we wrote a lot of custom
JavaScript code on our own.
And the glue of everything, so everything fits together, we
use required as a modular organization.
So this is a lot of content, and we try to keep the size of
data that needs to be transferred very small, so we
use separation of the content to static and dynamic content.
And we use cookieless domain for the static content.
And we try to compress everything and deliver it to
the browsers that support compression of content.
And next is some screen shots.
Maybe you already waited for it.
On the right side you see all the tags that
Oliver talked about.
The tags that were collected by our users.
That's Sulaymaniyah city.
And we try to use these tags wherever possible.
And on the left side or in the middle screen shot you can see
one example of this where we use it in the screen where the
restaurant administrator can create service zones for the
restaurant.
Oliver was talking about it.
And this is mainly for orientation, but also for
creating the service zones to know where you are on the map.
And we use these tags in a lot of other places.
When you order food, and for the deliverer, and so on.
And on the leftmost screen shot, you see one of our
restaurant partners.
That's the [INAUDIBLE] restaurant and menu.
And what you can see here is also that it mainly is using
jQuery mobile here, because we have a lot of information to
show on small screens.
So we use there, for example, collapsibles, for instance.
So you can click on the food category, and then it opens up
all the food there.
And then you click on it.
And there you can add different food items.
And everything's on one screen.
You don't have to switch around pages, something.
So everything's on one screen, and it fits very good into a
smartphone.
So that's something maybe you want to talk about again?
Thank you very much.
RALPH BENJAMIN MAGNUS: So I've already mentioned the task of
localization and providing the whole application also in
Arabic language.
So in the beginning, we started, if you wanted to do
the tagging campaign, we really had only two and a half
weeks to start to implement application to make the users
able to go around the city and tag important places.
So it was a short time.
So there for the localization, we simply had the text file
that we sent to our colleagues in Kurdistan, and
then we get it back.
But obviously, this is not manageable for application
that grows and grows and has to meet a lot of changes.
So we decided to have a localization back end, and
this screen shot is just depicting this fact.
So we see the application in English and then Arabic.
And you can also use application on the desktop.
For example here, for doing localization.
So the process of localization is totally embedded in the
application.
You can log in as translator and during the deployment of
the localization of translatable screens, I
extracted and provided in this interface that automatically
saves as you type all the necessary translations.
So this is just one example how we integrate the work flow
with the local stuff, as in Berlin.
And from this translation, all translation are pulled from
[INAUDIBLE] repositories to all different branches, so
that we have a consistent environment, at least
regarding the translation and the localization.
Just to give a little outlook at what will happen within the
next weeks and months, and let's have
a look how it works.
So Oliver already mentioned that we want to implement
something like a compass.
Why we want to do this?
Because the easy tags are very good to overcome, in the
beginning, the situation with lack of physical addresses.
But they're not so accurate.
So for the first, it's good to come near to the user or to
the customer who has ordered our food, but so the last 100
meters are important to find the customer.
And we want to implement such a compass where the delivery
boy can see which direction he has to
go to find the customer.
And it should be very easy.
Without maps, without street names, just during driving he
should be able to look there, and not to read there and
drive in this direction against the wall.
He should be able to have a quick look where to go.
And then we also want to reuse the tags that we have
collected so that not only is the user is displayed, or not
only is the customers displayed, but also tags that
are nearby that can help the delivery boy to get an
orientation of where he is, where he has to go.
So right of Canyon Hotel, for example.
Then one other thing is, of course, we want to allow the
restaurants to customize the menus to give it a
unique look and feel.
And maybe later to run independent
instances of Easy Bites.
But one of our main focuses is really to extend the work with
the tags to extend to overcome the absence of physical
addresses, to find different approaches to provide good
orientation.
And we have many ideas, and we are currently working on this.
We're doing research on this in cooperation with the
government, and so on.
So there will come a lot of new changes in the
next weeks for us.
But we are looking forward for the changes, because we are
prepared with our architecture and with
the application design.
So this is, I think, the most important things that we will
do in the next weeks, months.
We have collected some facts where we can talk about.
It sums up everything a little bit.
Which technologies we have used, what we have integrated,
which services we are using, and so on, I think that's all
that you want to say in the short introduction.
I think everything else we can do in the discussion part.
OLIVER FRANCIS KOCH: That wasn't that short.
Sorry about that.
Obviously, we couldn't give you guys the whole picture.
There's a lot missing.
We have a feedback system where you rate your order, but
it's all very basic.
But anyone have any questions?
Can we hand this around, or how do you want to do it?
MALE SPEAKER: You could repeat the question.
OLIVER FRANCIS KOCH: Oh, OK.
AUDIENCE: Have you got a sense of what standard devices
people normally use, like what kind of phones?
Is there a part of the application which is also on
the desktop or--
OLIVER FRANCIS KOCH: So the question was whether we have a
sense of what mobile devices are currently being used, and
if also the application is
accessed via desktop computers.
Remember right at the beginning, I told you
something about this MVP, where we
placed it on the website?
This cartoon.
When people signed up, we asked them to-- they could
choose four different categories.
Are you using Android?
Are you using iOS?
Are you using Windows Mobile, and so on?
So there we aggregated a lot of data.
And back then, our question was are we going native, or
are we not going native?
And I know Ralph had said something about the iPhone
being very, let's say, present, but that was the
first impression.
After we had that running for two or three weeks, we saw
there was a balance between Android and iPhone.
And once in a while, you had Windows Mobile and BlackBerry
slipped in there.
So I'd say right now we have about 40% iPhone, 40% Android,
and 10, 10 Windows and BlackBerry.
Regarding desktop, it's actually funny that you're
asking, because we have measured over the last two
weeks that--
basically, when the system cannot position you, we had a
fallback, meaning, where is it supposed to
deliver the things to?
We realized that a lot of people weren't using this
fallback, meaning a lot of people were getting this
message that they couldn't be located.
So we tracked this back, and people were using it through
their desktop, which was also part of it.
But the issue was that with the desktop computers, we used
wireless LAN to position them.
And the problem with Google Chrome, but also with other
browsers, obviously, that you get little pop-up to allow it,
and people don't really read it.
So they're just stuck there, and so they close their
browser at some time and just order via phone.
So that's something that we need to work on, too.
But definitely, those two channels are--
AUDIENCE: So at the restaurant, [INAUDIBLE]
OLIVER FRANCIS KOCH: The restaurant has its own view.
I think you saw parts of it.
Basically, most of them use a point of sales, and in Iraq, I
think like 95% run on Windows.
But it doesn't really matter for us.
We open up a little browser window next to it, and I don't
know if Ralph can simulate it.
They basically, whenever an order comes in, they get this
really, really annoying sound, which keeps going until you
read the order.
So that's how it works on their end.
AUDIENCE: With the menu, my question is basically, how do
they customize their menu on the app?
Does it go back to the point of sale?
OLIVER FRANCIS KOCH: No, it's not integrated to the point of
sale, because it's just so heterogeneous.
People are using own systems, and so that's not
feasible for us.
So they basically have a manager interface, and they
have an approver interface.
The approver is basically the cashier that sees the
orders, and so on.
The manager has the option of editing the menu, and so on.
And it's all--
AUDIENCE: On the app, or--
OLIVER FRANCIS KOCH: On the app.
On the app.
AUDIENCE: They open on their phone?
OLIVER FRANCIS KOCH: No, no.
They basically open up on a desktop computer.
And I think Ralph can just show you while I'm talking.
RALPH BENJAMIN MAGNUS: Just keep talking.
OLIVER FRANCIS KOCH: I'll just keep talking.
So yeah.
They open up on their computer,
have a little screen.
And they have the same view as here, just that they have a
lot more options of editing the menu, deleting orders,
setting delivery radiuses, and so on.
AUDIENCE: So if the restaurant owner doesn't have a PC,
[INAUDIBLE]
OLIVER FRANCIS KOCH: You could say it like that, yes.
But the issue is--
the way they operate right now is someone approves the order
and then they print out a sheet of paper.
On that sheet of paper we have a zoomed in view of the
location where it's going to, and a zoomed out view.
So you see the tags that are exactly next to the location,
and some that in a different perspective.
Because some restaurants, let's say maybe 80% of them
are like, OK, I'm not going to give my delivery guy a phone.
No way.
He's going to be gone.
So they're like, OK, what can you offer us?
And we're like, OK, we can work with this sheet of paper.
And that's basically what we developed with them together.
So they have a printer.
They print out the thing, the customer signs it at the
bottom, and that's it.
AUDIENCE: Do you guys have any issue that people were
complaining, the business owners were complaining that I
don't have a PC, so I can not customize my menu?
OLIVER FRANCIS KOCH: Well if you think about it, the
restaurants--
MALE SPEAKER: Keep repeating the question, because there
are people--
OLIVER FRANCIS KOCH: Oh, sorry about that.
So the question was, is it an issue that a lot of restaurant
owners don't have any computers?
Right now, the restaurants we're targeting, really, all
of them have a point of sales.
The only issue is maybe the internet.
But there's a thing called a HiMax, which is basically
something like 3G here for us, and they buy this little
adapter, pay $10 a month, and that's it.
And it doesn't take a lot of convincing on that side.
And strategically for us, and this is a very important
point, for us, this whole thing has a
large network effect.
Two groups, the users and the restaurants.
There are two or three restaurants, if you have them
on board, a lot of people join, and a lot
of restaurants join.
And the restaurants join if a lot of people join.
And it's a cycle.
And I think about two weeks ago we really broke that
cycle, because there's a restaurant called B to B,
which is in a Erbil.
Really, everyone orders there.
They have 200 or 300 deliveries a day.
Meaning you tell another restaurant, hey,
B to B is on it.
Oh, really?
OK, I'm in on it, too.
So that's basically how it works.
And I think we got to that point now.
RALPH BENJAMIN MAGNUS: And in addition to this, especially
in the beginning, we also assisted the restaurants with
importing their menus.
And we give lots of assistance.
OLIVER FRANCIS KOCH: Well, the issue is also that--
and that was training our sales team, the
people that go out.
We had limited resources.
So you're going to probably laugh if I tell you that what
we did is we got some students that really liked the idea and
told them, hey, listen.
Go around and start asking restaurants, hey, why are you
not using this?
And basically, go to the restaurants, and if you seal
the contract, we'll give you 50% of our revenue for the
next six months, which is achieved through that.
And that's working wonders, because we have really limited
capacity regarding our resources.
That's working around it.
And one of the main things training them--
maybe also a cultural aspect--
is that we don't even show the people the manager interface.
We go there first, and even while they're using it, we
just show them the approver interface.
Putting in the menu the first time, we do it.
Because they're just going to be so over overwhelmed by the
complexity.
They'll be like, ahh, no, I don't know.
So right now, it's three buttons.
We tell them, hey, listen, it's exactly three clicks, and
you've got your order.
And they're all totally happy.
Then they come back to us.
That's actually what has been happening.
They come to asking us, hey, is there something to edit the
menu and blah?
We're like, yeah, there's this, this, and this.
Oh, really?
And there's reporting, too?
Great.
So that's basically--
it comes from their side and at their pace, and that makes
it a lot easier for us to go along.
RALPH BENJAMIN MAGNUS: Thank you so much.
That was what I wanted to say with previous sentence.
AUDIENCE: I have two questions.
First, it seems you're doing cash on delivery.
There is no other payment, no other way?
And the second thing, do you, from your experience--
I know it is still early, but-- do you think users know
already to trust ordering something online, even if
[INAUDIBLE]?
OLIVER FRANCIS KOCH: Well, first thing, cash on delivery.
The question was, are we doing cash on delivery?
Yeah, there are no mobile payments, for whatever reason.
A lot of conspiracy theories there, but
let's not go into that.
The second question was whether there's a trust issue
within the local community.
I think--
I was there since July.
And monitoring what's been happening in Erbil and
Sulaymaniyah, especially regarding the engagement of
technology in day-to-day life, they're really opening up
towards that.
And like I said, a lot of people are already using the
phone to order food.
And the amount of those people is actually increasing.
And if we take a look at, really, our primary customer
segments, at the beginning it was really focused on students
who obviously like to engage with new technologies.
But at this stage, now, it's a lot of
private sector employees.
Companies like big telecom providers like Korek or so,
where they're just sitting in their cubicles, and they order
their lunch.
And for them it's just amazing.
It's their highlight of the day, just using the app to
order something.
Because--
AUDIENCE: Do you mean they are using the phone?
OLIVER FRANCIS KOCH: Yeah, they're already using it via
phone, and they're really annoyed by that process.
So within our marketing campaign, and that was really,
I think, that's the feedback we got.
It was a really important factor.
We put at the beginning made on Germany everywhere.
And they're like, OK, Germany sounds good.
We trust Germany.
So that made a big difference, too.
So it's a question of how you position
yourself in the market.
I think it's very sensitive to this.
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: How do we handle the quality
expectations?
Now, like I said, we couldn't show you the whole picture.
What basically happens, after you're done ordering--
and I hope--
I don't know if Ralph is actually working--
OK.
And the delivery guy went back to his restaurant and said,
OK, the order was delivered.
You would get a pop-up screen next time you log in, and you
have to rate the delivery service regarding time,
regarding food quality, and service quality.
This feedback always goes back to us, first, but also to the
restaurants on a regular basis, for numerous reasons.
We believe that it's a point we have definitely considered,
that bad service regarding the restaurant will reflect
negatively upon us.
So we have a responsibility of engaging there.
I'll just give you one example of how we would engage.
A lot of restaurants at the beginning were
like, you know what?
I'm going to take entire Erbil as delivery radius.
And I'm like, for me with my car, and let's say not rush
hour, it would still take me over an hour to get there and
food would be cold, and so on.
So what we try to do is we actually issue
recommendations then.
If we see that the delivery time is really a lot more than
average, we would go there and tell them, hey, listen, the
way you defined your service zones is just not working out.
And I think we have the leverage there, to be honest.
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: Yes, of course.
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: It's a mix.
Two interactive components that we have.
The first thing is on every order you place as a customer,
you can put in notes, which are retrieved by the
restaurant.
And the restaurant obviously can reject your order, approve
your order, but it always has the option of responding.
It's like integrated chat, if you want.
That's one interactive component where you have some
human components, if you want.
The second part is we do display the restaurant
telephone number.
The third point is also that what we try to monitor very
closely is how many orders have failed?
What were the reasons why they failed?
We also try to monitor very closely this dialogue between
the people.
If your order was rejected, why was it rejected?
But these are obviously issues that really become very
critical when we scale, when we grow.
And yes, we haven't reached that point yet.
And I hope we do reach it at some point, because
that would be good.
RALPH BENJAMIN MAGNUS: One addition to this is that we
also track if the restaurants are using the application when
they're intended to use it.
So they can set up constraints of opening hours, and we can
monitor if they're really having open this site on the
left, where you see the incoming orders.
If they are offline when they should be open, Oliver and
Christoph and Mohammed will get on the email.
And Mohammed has to call them and say, look,
switch on your computer.
You guys should work.
OLIVER FRANCIS KOCH: You're laughing.
It actually happens.
And they're like, how do you know?
Well.
You hacked into my computer, or something.
No.
And it's an evolving process.
We didn't have all of this from the start.
It actually happened I ordered one day, and
then nothing happened.
And he's like, OK, let's implement this.
AUDIENCE: Two questions, one from a person dialing in, and
one from myself.
The question from me is, going back to the business model,
what's the difference between a commission and a $1 charge
for ordering?
Practical terms and cultural terms.
OLIVER FRANCIS KOCH: Cultural terms, I guess--
I don't want to offend anyone, so I got to be
careful how I say it.
If we take a look at the average order volume.
The minimum order value of most restaurants is
about $10 to $12.
And for them, it's obviously--
I think it's an economic motive.
They basically see, OK, no matter what we do, we're
providing the same service.
So why would we get be getting more money if someone's
ordering more?
That's their perspective.
I think it's cultural.
At the same time, it's also practical.
Because if you take a look at it culturally,
usually it's families.
Families, the normal household size would be about five to
six people.
So it wouldn't just be one person ordering, it would be a
person ordering for five people.
So we're talking of orders of $50 or $60.
So for them, they're basically saying, OK, so you're not
doing a lot more when one person is ordering or five
people are ordering, so I just want to stick with this fixed
fee of the minimum amount, which is
on average $1, exactly.
AUDIENCE: My question is about the [INAUDIBLE].
OLIVER FRANCIS KOCH: Yeah, you want to run another
competition with us?
That's cool.
You can sponsor Nexus Ones, or something.
No.
So what was the design of the competition?
Well, obviously what we did is--
people didn't even know what tagging was.
So we invested a lot of time in educating
people in that regard.
Facebook is a very popular channel.
It's probably one of the only channels that you can use to
target the audience that we were trying to target.
So what we did during the set up, we said we're going to
focus on Sulaymaniyah and Erbil, because there's no use
in just tagging and not validating the data.
So we had a couple of students who were freelancers who were
looking into the tags, saying, does this make sense?
Is this realistic?
And they would then approved the tags.
Because that was also an adjustment that we had.
The first couple of tags, we had the first day one person
who did, I think, 150 tags.
So we were like, oh, this is going like crazy.
And then we realized she was tagging
everything from her house.
So we were really adjusting that.
And he basically fixed that you can tag things in a
certain radius, and so on.
So it's an evolving process.
In total, we ran the campaign, I think, about
five weeks in total.
We have accumulated in Sulaymaniyah about 350 tags, I
think, and in Erbil a little less, about 200.
And they're a little wide scattered,
so we're lucky there.
AUDIENCE: Is it private data?
OLIVER FRANCIS KOCH: Up until now, if you guys
have use for it, sure.
I'm saying that.
But we invested some money into it.
So I think we're open minded, but right now, it is private.
But again, if there's use, definitely we're open for
discussions.
We're not hogging.
AUDIENCE: So you mentioned about the internet
penetration, what the percentage may be.
But what about smartphone penetration?
OLIVER FRANCIS KOCH: We have data on mobile phone
penetration.
I'll give you that first.
And it's going to stun you, because it's
90% in whole Iraq.
But then, you have to look at--
and this is, again, this is more a qualitative assessment
that I can give you, going around and telling
you what I've seen.
Is that especially in the target segment that we're
focusing on, 18 to 35-year-olds, all the people
are using smartphones, especially
in the Kurdish region.
And we also have people that come from Baghdad, that see
Baghdad, and I hardly--
anyone that has a cheap old Nokia phone in that age group,
always has his iPhone along.
And this culture is basically all about logging into
Facebook at every location that you go to.
It's cool, but it's different from here.
So what we can say on that end right now is we have no hard
data on that.
But we basically--
it's more based on our qualitative assumptions that
along with the internet penetration, a
lot of it is mobile.
And I think a next step for us to narrow that down would also
be to see in how far the internet penetration, which is
already measured through Facebook, is to see and how
far is it also being used mobile?
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: I'll give that to my--
AUDIENCE: How do you maintain your 99.9% of time?
RALPH BENJAMIN MAGNUS: How we maintain this?
Right now, we sit there and pray all the day.
No.
We have running some monitoring services, of
course, that are notifying us when something happens.
So we have serial instances running of this service, but
only one is known for the live system.
Others are reserved for testing and for localization.
For now, we have no redundancy.
But this will change if there will be some money and we have
the ability to invest a little bit more in the
infrastructure.
But for now it's working well.
And we can react very fast, because, of course, monitoring
from external services.
So we have a chance to maybe react within some minutes.
[INAUDIBLE]
AUDIENCE: [INAUDIBLE]?
OLIVER FRANCIS KOCH: Well, up until now, we had a couple of
local companies, too, that showed interest in sponsoring
this, and so on.
We're still bootstrapping, because, like I said, for us,
[? Karoge ?]
is our pilot market.
Our goal is actually to expand to Baghdad, and so on.
But if there are any options of acquiring funds in another
way, we probably would do it.
But we haven't gotten to the stage where we are actively
seeking seed funding or anything to push this or boost
the business.
It's also, we still feel that we are really learning a lot
regarding the changes.
Just taking a look at the last months, we're not at the--
I mean you reach the point where your learnings--
obviously, at the beginning, your learning
curve is very steep.
At some point it flattens, and we have not completely reached
that point yet.
When we do, and we get our internal responsibilities and
all settled, I think that's the stage when we would start
seeking external funding.
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: Totally.
Currently, one of the biggest issues from the restaurants
point of view was, OK, so this is internet and people are
ordering, and couldn't people just fake order and stuff?
So what we did is, when you register, you have to give
your phone number and you get a code.
Now, you're going to laugh, but we actually took this from
Eric Ries' Lean Startup idea.
We did it really manually.
I mean, we have someone sitting there sending out text
messages, this is your activation code.
So at some point, yes, that's definitely something we have
to look into, or want to look into.
AUDIENCE: You mean ordering.
AUDIENCE: Yeah, that was my question.
Just it provides another--
OLIVER FRANCIS KOCH: I think that when we start integrating
SMS gateways and so on, that would be something that we
could approach then, too, obviously.
AUDIENCE: [INAUDIBLE]
OLIVER FRANCIS KOCH: Like I said, we had a platform.
We were conducting these startup trainings.
And a lot of things in that community are--
word of mouth is really, really, really, really big.
So what we did is, in one of the trainings we did a user
acceptance test.
We basically showed a video--
we have it here, too-- but which basically explained the
idea of Easy Bites.
Kind of like that Dropbox video.
You might have seen it.
And we told people, hey, OK, you know what?
We're going to post on our Facebook page.
If you like it, sign up and share it.
And that's how the whole thing got started, actually.
And the next stage then was to build a website where we put
this MVP on.
And I think up until today, for some Facebook and flyer
campaigns we spent maybe $1,500.
That was more from our own pockets here and there.
AUDIENCE: OK.
I think that's it for now.
Thank you.
OLIVER FRANCIS KOCH: Thank you.
RALPH BENJAMIN MAGNUS: Maybe just I can just show this
small use case.
I don't know.
I can't remember why I wanted to show, but maybe it's
interesting.
Here on the left side, I'm logged in as an approver.
This is also a role that has come up from our observations
that not everything has to be done by a manager with all
privileges.
We have just one guy sitting there that approves
deliveries.
OLIVER FRANCIS KOCH: Cashier, basically.
RALPH BENJAMIN MAGNUS: Cashier, exactly.
So he has very limited possibilities.
And this button is also one observation.
We have to have the ability to switch off delivery
immediately, if maybe the delivery guy leaves for lunch,
or something unplanned happens.
So this overrides the schedule of opening hours.
So what happens if I've placed my--
I put my position now to Erbil in a location
where we have delivery.
I choose my own category of desire, and I choose to order
from this restaurant.
What I get here is their menu.
Well, my son loves pizza, so I order some pizza.
Let's try it with chicken, he like.
I take also one.
And he can give some information.
OLIVER FRANCIS KOCH: He can comment on this, because the
complexity of the whole process gets more complicated
if we have a shopping cart.
So we kept in this very easy, straightforward design, where
you see at the bottom here, you just have a listing of
what you have and you can cancel it out.
The cost is, obviously, it's not as flexible, but it keeps
it in a very, very streamlined, straight process.
RALPH BENJAMIN MAGNUS: So now the customer orders.
It comes up with nice sound, and you see the
overview here of orders.
To cancel this, you have to go to the order, and you get
overview, including the customer notes and telephone
number to call the customer, maybe to call something is
out, or do you want to change something.
You can edit then if something should be changed on request
of the user.
So then we had--
after one month of implementing, we
had a list of seven?
Seven states that the delivery can have.
It was very detailed, and very cool.
You can follow each step.
This totally failed, because nobody was
able to handle this.
Now we come up with two states,
in progress or rejected.
So 30 minutes.
So the customer also has here some information.
So submit it, and the customer here should automatically get
an updated information that it's in progress.
So here we use our JSON interface to pull the status
of the delivery.
The cashier gets something to print for the delivery guy--
the route and screen shot of position, very close screen of
the position of the customer where you see, OK, he's close
to this tag.
That's one of the tags that has been submitted by our
customers and users.
And then you can set it to delivered.
And the user now is blocked for a new order until he has
rated the order.
So in this way, we get a feedback that delivery has
taken place, and we have to get our
dollar from the delivery.
So he has to give this.
Our observation is that this actually is done when the user
wants to order new food three days later.
So it's never done immediately, only we have to
wait three days, and then it's completed, the order.
And then it can continue.
OLIVER FRANCIS KOCH: This process was basically--
the way we changed it.
He said we had a seven step process.
It was really complex.
And just speaking to the restaurant owners, they're
like, I don't get it, and my cashier
definitely won't get it.
So what can you do to reduce this?
And then we came up with, OK, you know what?
It's just in progress, rejected.
And then after that, you print it out, and it's delivered or
delivered failed.
And that has really served us a lot, because it has really
reduced the complexity of the whole application.
Because you can explain it to anyone in five seconds, and
they get the idea.
RALPH BENJAMIN MAGNUS: Maybe just to close it, because
there was a question.
I'm just now take the role of the manager of this restaurant
that I ordered right now.
And the manager has the possibility to maintain the
whole menu.
He can arrange the menu, the categories, the options.
So different sizes he can define for
pizza and for drinks.
And so everything can be done directly in the application.
I think that's all in the short
time that we can present.
MALE SPEAKER: Thank you again for this event.