Live Office Hours with Adora Cheung and Avichal Garg Video Lecture - Startup

Hey everyone, for todays live office hours
we're joined by Adora and Avichal and
I'll leave it to them to
introduce themselves.
Adora Cheung.
I am a partner at Y Combinator.
Previously I had a, I was a founder
CEO of a startup called Homejoy.
>> Hey, Avichal Garg, I am a YC expert.
I was was recently director
of product at Facebook.
Done a couple startups,
sold a couple, invested in a bunch.
[BLANK AUDIO]
>> Angela.
>> Hello, Adora.
Nice to meet you.
>> Avichal, good to meet you.
>> Adora.
Nice to meet you.
>> Avichal, good to meet you.
>> Cool, welcome, so, tell us what you do.
Hi, we're working on Appmission.
It's a web application
that offers high quality,
personalized college admission advice for
high school student.
>> So is it a, is there not a person
on the other end, or how does the?
Can you give the most common
use case of how people use it?
>> So a student can sign up for free,
and then they will create a profile.
>> The profile includes the academic
information, including the GPA it
has scores, and also extracurricular
activities, and the college goals.
For example, the colleges that
they want to get into, and
also the majors they are interested in.
So once they enter the data,
we generate algorithm to-.
We use our algorithm to generate
a personalized recommendations
to them to help them reach their goals.
>> Gotcha, so who's using this or
who do you want to use this?
>> The high school students and
their parents.
>> Who more specifically?
>> For those who cannot afford
private college counseling
because that is the price range is
from 4,000 to $40,000 per package.
>> I see.
>> And why would they use this?
>> Because in the U.S. public high school,
one counselor serves 491 students, and
each student only get 38 minutes
of college advising per year.
So we want to fill in the gap so
they can get access and
we think that every student deserve
the help they need to get into college.
>> Is there one particular thing
that they really need help on?
Because one of the things I
think a lot of people only buy
a lot of books about is essays,
because the application itself is pretty
objective diifferent
objective in standard, right?
So do you find there's a specific feature
that people are really looking for
and will actually pay for?
>> Yeah, that is interesting.
So when we first started,
we just came pretty broad.
We talked about our essays,
we talked about college selections.
But recently,
when we talked to some of the users,
we found out that they want
one on one interactions.
And specifically on college selection and
the essay.
>> So are you building a marketplace then-
>> [LAUGH]
>> Of people?
>> So when we first started,
which we just kind of did it on our own.
The back end is us but actually we talked
to Anton our mentor on this, our school
and we also got some feedback from users
saying they want one on one interactions.
So last weekend our team spent 30 hours
to build and launch that marketplace.
>> And how's that going?
>> It's going quicker than I expected.
So we, for the past three days we
had about 800 people visit the site.
And then we had 60,
about 60 advisors signed up.
So right now, because of marketplace,
we focus on the supply side first.
>> Okay, so you have 60 advisors.
And so you're focused on acquiring those-.
Do you think that's enough?
That seems like enough, because -.
>> It's Excuse me?
>> I don't think you can hear me,
can you hear me?
>> Hear me and [INAUDIBLE]?
>> There you go, this is not on.
>> [INAUDIBLE]
>> Are we doing okay?
I can yell also.
[INAUDIBLE]
>> Yeah, yeah.
>> [INAUDIBLE]
>> Does this make it better?
>> It may not have been on before.
Is that better?
>> Are we good?
Are we good?
Okay, cool, sorry about that.
So acquisition.
So you've got 60 advisors.
It looks like you're ready to go then,
right?
>> Yeah, we're ready to go.
But then we, right now our next step
is to review those advisors and
make sure their profile is nice and that
they are high quality of counsel that will
have specialized in certain areas so that
we can use those to attract the students.
>> Why is now the right time to do this?
LIke this could have
been done five years ago.
Has something changed in the world
that makes this possible now?
>> I think the industry has been growing
slower than other
consumer-based technology.
So I think right now is a good time.
Because the parents and student,
they are getting more and
more tools in their school
using technology to learn.
Like personalized learning and
collaborative learning.
So I think this is the space where
it has not been fully utilized, yes.
>> Why do you think that is?
>> I think the reason why is.
People usually,
they focus really on tutoring.
They've put a lot of those energy
onto getting good grades, and
then they feel like they
know exactly how they
can get into college just based
on their friends and family.
So they rely a lot of those
personal advice from the friends or
family or
whoever would have gone through that, but
we want to give them
the professional side of things.
>> It seems like the amount of information
that's out there is out there and
so peeople are trying to probably
differentiate themselves because there's
the internet and I can get that same
access as a person a or person b.
And so it makes sense that
you need somebody else to tell you how to
make it better versus just reading stuff.
Bu okay, so I want to focus on,
it seems like your biggest challenge
right now is going to be matching right?
>> Mm-hm.
>> So you have 60 advisors,
it sounds like you believe they're high
quality right now to start matching them?
>> Some of them.
>> So how are you going to, or
what are your top two pre-determinants
of bringing on these advisors?
>> So definitely one of
the things that we're looking for
is, whether the counselor or
the advisor is certified or licensed.
Again, we just launched a few days ago.
What we are going to do is to focus on
those profiles that are existing right now
and talk with the actual advisors to see
if they have the credentials, right?
>> Are you advisors yourselves?
>> I am.
>> Okay, so you're on the site?
>> Mm-hm.
>> And have you-.
I mean one thing would be to just match
people to you, to figure that piece.
>> Right.
So I already start doing that.
I had one person, which I met in person,
because they basically just
wanted to meet with me in person.
That's how we got.
We think that,
that this something that will be valuable.
But already start putting myself
out there into, on the supply side.
>> So there's a whole set of product
risk here which I think you guys
will work through.
The thing I keep coming back to is
actually the market risk on this.
because my first startup was in education,
so I know a ton about this space actually.
And historically the challenge has
been that they way that the average
person in the US thinks about things like
college admissions is actually pretty
different from people who are in this room
probably think about college education.
And fundamentally a lot of that comes
down to do you think about education
as an investment or an expenditure and
this happens in a lot of markets
where you kind of build for yourself.
And then you realize the market doesn't
actually behave the way that you behave.
And so there's a leap that sometimes
companies aren't able to make.
And so
that's why I was sort of probing on it.
I was like has the market
actually underneath this changed?
Like have you actually validated that
people want the same things you want?
And that that market is large enough
to actually support a business.
>> Well actually,
there is reason change that I think,
will give us a opportunity
which is a new administration.
Because we are talking about funding
cut into public school system.
>> Uh-huh.
>> So we would expect that the resources
that they have will be less and less.
And I think that is an opportunity to
educate the parents and student and
to promote.
The needs of having the college advising.
>> Got you, but
they don't currently get it today?
>> Right.
They get very minimal.
>> What resources do people have?
Because there's books,
there used to be sites like Volt.com
which I don't think gives advice anymore.
>> Yeah.
>> [LAUGH] So
what are they using right now?
>> Right.
So there are a lot of free sites
out there like college for,
they have check.
They have different websites that tells
them the details of the colleges that
probably the application rates and
other stuff.
But those are not personalized for them.
And when we did an interview
with the users,
they found it very overwhelming to go
through all these different sites.
>> Who were the people that
you were interviewing?
Where did you find those?
>> We found them from our neighborhood.
Found them from our friends kids and
basically people that we know.
>> Yeah.
I keep coming back to this.
I think specifically, this is one of the
biggest risks in the education market is
that the mass market is going to behave
differently than the people that you know.
And so actually validating by talking to
those people what their needs are and
what you want to build for
those people and what they want is
actually pretty critical early on.
Because people from your neighborhood
probably think about education very
similar to the way you do.
Or your friend's kids probably think
about education very similarly
to the way you do.
>> That's why every since we launched,
we try to iterate different ideas onto
the site to see which one
catch more of the attention.
So far I think we were able
to identify that one on one,
personalized interaction is important.
But I think the next step we wanted to
drill down into something more specific,
so that we can just
focus on that one thing.
>> Yeah.
Have you guys considered the international
markets at all?
Is this just a U.S.
problem, or is this a global problem?
>> We thought about it.
There is some leads that we talked to.
There's definitely the potential there for
international markets.
But the platform that we have built out so
far is really targeted to U.S., but
that doesn't really limit us
on expanding to more markets.
>> Have you thought about
going international first,
as opposed to U.S. first?
>> So our target was basically
States first, that was the focus.
And in fact our website is admission.us,
so it's more like, you know.
But definitely again, we do know
that there are people out there,
outside of the United States who
want to study here in the States and
they need that information as well.
We talked to advisors, as well,
from outside of the country, and
there is the need, but
we don't really have an idea,
you know let's target those markets yet.
>> How many users are on
your site right now?
>> We have about 200 students on site.
>> 200 students, and what's the common
element amongst all of them?
>> So we have 40% of those are senior,
so that tells me that they're pretty,
they wait until pretty
last minute to do this.
And then 25% the junior.
A majority of them are focusing on
computer science, and engineering,
math type, and business majors.
>> It sounds like you might want to,
and I think this is kind of what Avichal
was getting at also, is try to figure out
what is that core thing that they
really need, versus just like.
Right now it seems a little abstract of,
I'm going to help you find a college, but
it might be something as simple as
the essay, like I need help on the essay.
And just focus on that one thing, and
drive it home verses being a little bit,
doing some many things at once.
>> So our mission is to basically
make sure that we provide the same
opportunities for all students.
That chance to go into college.
And we understand that definitely
within maybe California,
maybe in bigger cities that
students are more educated.
They have the background of
what's the path to college.
And beyond that what we are trying to do,
again you know going back to mission,
we want to make sure that outside
of these group of students,
people from outside California also have
the same access to the same information.
I mean if you think about or
actually based on our survey or
talking to users we know that
a lot of students actually think
they might not even have the chance
to into colleges because it
might not be affordable to them, right?
But there is definitely alternatives for
them like financial way and
things like that.
That's the tier that
we're trying to target,
to make sure that they have
access to that information.
So that they know they at least have
the chance to apply to college, and
maybe qualify for financial aid and path.
>> I'm just going to echo what
I think Adora's saying is,
I love what you said by the way.
Like that's an awesome mission, right?
And so there's, when you think about
a company there's like the mission,
there's the vision.
Like when we achieve the mission
here's what the world will look like.
There's the strategy,
which is like, here's how we're going to,
in the fewest number of words,
how we're going to go about achieving our
mission, and then there's the tactics.
There's the things that
you do week over week in
pursuit of your strategy to validate
whether that strategy is true or not.
I think the advice that Adora's
giving is more like strategic advise.
Where it's like okay, great, that's
an awesome mission but where do you start?
Like what is the first
value that you create for
people and then that lets you create that
second value, and the third value, and
then you've created all of the value for
that population of people.
But where do you start?
>> Yeah, if you think about if
you're going to apply to college,
there's many steps in that and
you can't tackle every single step.
So, find the one in which people
are having the most issue with or
they need the most help and
which they will actually pay for as well,
because you're trying, obviously,
build a business, right?
>> Right.
>> And, it's not just product strategy.
There's also the go to market strategy,
which is, that's a lot of people, right?
How many people, is it like 50% of
people even that take the SAT in the US?
Right?
It's like half
the people don't even take the SAT, right?
So you're talking about large numbers
of people at every step of the funnel.
And so who you target is also part of your
strategy, and who you target is going to
be directly tied to what you build for
those people as the first step.
>> Yeah, I think that's a great point and
when we first started there are like
many different angles we can target,
and we want to use the users' data and
we want to just ask the users to
drive us to which is the focus.
And I felt like because the early
acquisition is really hard for
the users so I don't have enough people to
actually give me feedback on which areas
they want to focus on.
Some early indication is
school selection and essay.
So these are the two so far that I can
identify from major is a science, math,
and business.
So those are three majors.
So we have some pieces of signals
that tell us which way we should go.
But I know, in my heart,
that going forward there is some,
one area that I need to focus on.
I feel like we don't have
enough data to get there yet.
>> Yeah, well, the challenge of data,
this goes back to
there is two ways to figure
out what you want to build,
whether it's your go to market plan,
your product is the top down.
Which is you go get a bunch of data.
Where are the opportunities?
How do you slice the market?
We're going to start here and
then we're going to go here and
then we're going to go here.
Or bottoms up which is you build
the thing, you put it out there,
people start using it.
They're like we want this.
We move forward.
And a lot of software actually bottoms up,
is great.
Because the early adopters indicate where
the world is headed as we build for
them and then the rest of
the world kind of catches up.
It's like social networking,
obviously is a great example of that.
The challenge is if the people that are
the early adopters of your product are not
representative of what
everybody else is going to do,
then you end up building a thing that
doesn't actually work for everybody else.
And that's, I think, part of the challenge
that you would have right now if you do
that is, I suspect the people that
are coming to you are not going to be
100% representative.
To me it sounds like
two different populations of
people that you're talking about.
The people who don't know about financial
aid and maybe are not applying to colleges
and not doing all this kind of stuff and
don't have the support, versus like
people who are organically finding
this and care about this stuff a lot.
Like to me it actually sounds like
different populations of people.
>> Okay, I think we have to finish up,
but just one last thing It seems
like a lot of people are coming to you for
advice anyway.
I might just start from there, and
just sit down with all these people And
just do it all day every single day for
the next, I don't know, few weeks.
And then, see if they're anything common
things that people are asking you.
Also, you have access to now 60 advisers.
You could also sit with them and
do the exact same thing.
It's totally not a scalable
thing to do in the future.
But, it seems like to figure out where
to start that's like one way to go about
doing it.
>> Okay.
Thank you.
[APPLAUSE]
>> Hey.
>> Have a seat.
>> Hi.
Adora?
>> Good to meet you.
>> Cool thanks for coming.
>> Thank you.
>> What are you guys working on?
>> So I'm Sid and I'm the CEO and
co-founder of Yembo.
At Yembo, we are bringing computer
vision technology to solve the price and
cost estimation problem for
the home services market,
specifically we are starting
with the moving industry.
So let's say you are consumer and
you want to move, and
you want to know how
much it's going to cost.
You take a video and our technology
will identify all the items, and then,
you get an upfront and accurate quote.
So this is much better than
calling multiple moving companies,
staying on the phone for about half an
hour each to get arbitrary ground codes,
or waiting at home for
on site estimator to come visit.
>> So is this an app?
>> So we-
>> Sure.
And we have multiple channels,
we did an NVP pilot with an app.
We've also been working on just
having it directly on the web page of
the moving company.
>> So say, I'm at home, it sounds like I
take the app and I'm taking pictures and
I'm sending it to you guys and then,
you kind of give me an estimate ASAP,
essentially, okay.
>> Correct, yeah,
and we also have a solution where
you can text it to a number.
>> Mm-hm.
>> But it's the same technology.
>> Okay, and
this is more accurate than the, or
what's the percent of accuracy do
you get versus someone coming in?
>> So
we are targeting the same accuracies.
So the way we are thinking about it
is today the estimation process takes
about a week for you to get a quote or
you get multiple quotes and
you don't really understand
why a quote is what it is.
So our goal is to give you a complete
breakdown on what it is, and we still have
some human labor in the background,
cleaning up the mistakes the AI is making.
So our goal is to reduce the human
time to less than five minutes.
>> Why is it so expensive,
or why is it so,
take a week to get a quote like
where moving take coming along?
>> So one main reason is if you are
scheduling an onsite estimator with it.
So it costs the moving company a lot,
more than $200 for a visit.
And you have to schedule
someone to come visit.
And these guys are able to
do 2 to 3 estimates a day.
So in summertime and it's busy,
it gets really backed up.
So most companies don't end up actually
sending an on-site estimator, and
they give price quotes over the phone.
Our research shows that 90% of this was
wrong, so consumers shop around and
that they don't really understand why
a companies quoting what they're quoting.
And in a lot of cases what happens
is there are these predatory
providers in the market who take
advantage of this price opaqueness and
they give you a low ball quote,
get your business, and
on the day of your move specifically,
they slap you with last minute fee hikes.
So we want to eliminate all of that.
>> So once you get the quote are you
also doing the moving yourself or
are you selling leads?
>> So, our strategies,
we're working with moving companies
to do the estimating for them.
So we are not a moving company, and
our customer is the moving company.
>> So you're a white label solution for
them.
>> Correct, yes.
>> Okay.
>> Why do you need computer vision for
this?
It seems like everything you described
to me sounds like process failures,
like transparency in a market place
failures, like it sounds like a bunch of,
well what does the computer
vision give you?
>> So our initial MVP that we deployed
doesn't have any computer vision.
People just take videos and
someone watches it in
the back end to give a quote.
That takes around 90 minutes per person so
and
also you need to hire and
scale human workforce to do that.
And there are a few other
companies that's doing this.
So our goal is, we eliminate like
majority of the work through using CV
and keep the human workload
to as less possible.
And from a product perspective
what we want to be also able
to deliver to consumers is a interactively
of them customizing their quote.
So let's say you have a $2,000 budget and
everything,
moving everything is maybe
going to cost $3,000.
Then you can decide, pick and
choose how you want to spend that money.
>> How do you intend to charge for this?
And do you have current
customers right now?
>> Yeah, so we signed our first deal for
a paid trial with a moving
company in San Diego.
There, we are doing this GPM solution
where it is $10 with no AI and
$50 with AI and
a we are talking to multiple
large growing companies from $20 million
annual revenue to a few billion dollars.
And we are in the final stages of
negotiations with some of them
on the timelines as well as
the features of a paid pilot.
So it's like enterprise level
contracts that we're going after.
So what does success look like for you?
Is it accuracy or
what are these folks looking from you?
>> So, two different things.
From a consumer standpoint,
they want to reduce the friction of,
they want to provide a better quality
of service because there's no way for
the good companies to
differentiate themselves right now.
So they are looking at it as
a better way to do customer service.
And also from a cost stand point,
since it costs more than $200 to send
an on-site estimator and it doesn't scale,
they're looking at ways to increase
their bottom line and their profits.
>> I wonder if there's a bigger
market in you being the initial,
people thinking you're the initial moving
company, or and you are selling the leads.
I don't know if you want to get
into the actual moving itself,
that seems quite different.
Just operationally a headache,
but Instead,
part of what you're doing
is also acquiring users and
getting that sales process going and
making it a great experience.
And if you do that really well, then,
why not just own the entire process?
The entire sale process, I guess.
>> That's something that we've considered,
but
moving has a lot more,
I guess intricacies and just the code.
So right now, we're just hyperfocused
on generating the best code possible.
That's something that we could
possibly do in the future.
And the second point is we're
not acquiring customers today.
The moving companies acquire
customers themselves.
So we are a technology that it's
either there on their website or
they direct consumers to our app.
>> Why do you charge more for
the AI and less for the people?
>> So with the $10 solution,
the assumption is if you're the moving
company, you watch the video and
provide the coach yourself so
>> So there's some human,
if you're the moving company, you have to
hire people to actually watch the video.
>> So
you're not hiring the people to do it.
>> On the $10, no.
>> I see.
>> Who is actually,
taking the non AI version?
>> Right now,
our AI is not completely ready yet.
So that's the product that
we are selling right now.
>> I see, but why not sell just the AI?
I mean, that's what you're going to be.
So why not sell the AI product now?
And you could still have
yourselves in the back.
Yeah, fake it 'til you make it, you know?
>> So that's a good question.
So two reasons.
The first one is do you want to train
RAI on real world people's homes?
You can imagine when you're moving your
house looks nothing like the beautiful
catalog pictures-
>> It seems like this is,
you need a lot of data, and
to get a lot of data really quickly, You
could just provide this blanket solution,
and all the moving companies
think they're getting this thing.
>> Yeah, and so-
>> And
you're providing the same amount
of quality either way, right?
>> Yeah, so we had actually hiring
annotators that are outsourced,
where as our AI is improving
their workload is reducing, so
that's something that we're doing.
>> So homes are very different, and which
is why it's probably such a big problem,
no one has figured out how to
standardize the quoting process.
How much data do you think you need before
you've gotten something that can just
run on its own?
>> So we are actively also web
scraping for data from Google Flickr,
so we have a few million
submissions already downloaded.
So the goal, again, with the AI is the
accuracy keeps increasing with time and
as more consumers use it.
And we have a fail-safe
mechanism of humans filling
in the gaps in the mean time.
>> What do you think your biggest or
what is your biggest challenge right now?
>> [LAUGH] That's a good question, so and
it's one thing I think startup school
has actually helped us out with.
So we are going after large
enterprise contracts, and
it's taking some time,
because our AI is not ready yet.
So we are looking at ways of
how to deploy solution while
simultaneously can monetize it, as well.
So in the meantime,
because of our mentor from startup school,
we are doing this GPM solution so
that we get our product out the door.
We did a non-paid pilot, so this is
the way to get some paying customers.
So anything that can reduce the sales
process, especially as we are going
after the enterprise customers,
would help us accelerate faster.
As well as, I guess,
we need to clone ourselves to create
a product that's much accurate
in a shorter period of time.
>> Have I got that right,
that it's generating leads, and then, so
what does your pipeline look like?
How many companies do you talk to a day?
>> Sure, so as Sid mentioned,
we're targeting, initially,
these larger enterprises,
because that allows us to scale faster.
So we found it's pretty easy to get
companies in the funnel when you explain
what you're doing.
It's the end of that, after you've shown
them the tech, you've shown them the demo,
and we're ready to sign for a paid pilot,
that a lot of folks are saying,
this is great,
I need it to do everything right now.
So there's a bit of a gap
between where we are today and
where we need to be where we
could close them a lot faster.
And that's kind of where we're coming
in with these creative ideas of
how do we monetize what we have today,
and then work towards that goal.
>> What is that gap, exactly?
So they want the full AI, and
it just produces a price quote.
>> Correct, yeah.
>> And do an [INAUDIBLE] list.
>> Yeah, so you can imagine that we're
detecting about 40 objects today, so
we need to scale that to 150 objects.
And because we're using video, we need
to track these objects across frames.
So we understand that this is one chair,
even though I'm seeing it for
from the 5th second to the 15th second.
So as we need to identify
the size of the object,
because that contributes to cost as well.
So there are multiple things
that we are doing, so
we are working simultaneously and
making progress in all of these.
>> And are they willing to sign LOI's
that say like, hey, if you had this,
we would give you $10,000 a month or-
>> Already signed.
>> I see, and so
how many of those LOI's do you have?
>> So we have half a dozen
of LOI's already signed, so
we still working on
the contract dollar amount.
With a few customers,
we are in the final stages of figuring
that amount actually is going to be.
>> And how are you pricing that?
So for the ones that you do have signed,
how have you priced it?
>> For the first year,
we are pitching a flat fee, so
a few $100,000, where they can use
our product as much as they want.
Where we actually work together with
them in refining the process of how
consumers actually use our
technology to get a quote.
>> I see, so the total contract size, or
each contract is worth a couple 100,000?
>> So we're looking at signing a maximum
of up to three, in the first year, for
a few $100,000.
>> I see, so couldn't you just,
to Dora's earlier point,
couldn't you just hire two
people to do this full time?
And go get a couple $100,000 from
these guys and say, hey, we have AI,
that's as good as a human because
you literally you have a human?
>> That's what we're doing, I mean,
that's out goal once we sign this.
But we also want more
turnaround time of a few hours.
So that's something that we're
working with actually, so
that's a very good point.
>> What is your turn around time?
I guess your optimizing
this at this point.
>> Yeah, so it depends, so we can have
our turnaround time as less as one hour.
It just means that we need to make sure
that we have someone on call 24 hours
a day, so when when RDA spits out
the analysis, they're able to clean that.
>> You might actually just want to figure
out what the labor cost there, for now,
as you're billing up to data.
And then just price accordingly.
So right now you're charging $10 or
something $50.
>> Yeah.
>> $50, and right now,
due to the labor costs,
because you don't have all the data,
it's not accurate enough,
it might be $100 cost to you.
So why not just charge something
a little bit above that?
It will knock out a number of
people who won't pay that much, but
at least it'll get you
going with real customers.
>> So, the large enterprise customers
we are pitching are flat rate.
So the smaller guys pitching the $10 and
$50,
just to get more data through our pipe
line, but maybe that is a good idea.
>> Why are you pitching flat
rate with the big customers?
>> Because this is very new to them, and
we want to make sure they are incentivized
to use it as much as possible.
And pitching on a per use fee,
we feel like,
maybe-
>> Do they already have estimators?
Because you are replacing them,
in some sense, down the road, right?
>> So I don't think, initially, the target
is, let's say, you have a five bedroom or
six bedroom house, they may still
want to send an estimator over.
But there's sort of a significant,
maybe 75% of the cases where they
don't really send an estimator.
So initially, based on our conversations,
they may start using as far as that
specific segment of the market or before.
>> This may getting a little
into the leads, but
one of the advantages of sending in
estimators is they're also a sales person,
and so they close the sale for them.
And so it might be that,
especially if you're going into a huge
house that could be a lot of profit for
them, they wouldn't even want to just
send them off to somewhere else.
>> That's correct.
>> So maybe is you also,
this is a tool for that estimator, right?
To go faster, and
they can get to more houses, and
they-
>> Yeah,
that's one of the options that we're
working with, with one of these companies.
>> Great, okay, got it.
>> So what does this become?
Suppose you're successful, you get
a bunch of contracts, then what happens?
>> So again, we started with moving,
but we've also found opportunities in
other parallel verticals, like painting
and home renovations and, so on.
So where the data that we're collecting
and the AI that we're treating,
will help us get to a solution faster.
So our goal is solve the problem for
one industry and
go horizontal to the second industry and,
so on.
>> It seems something like
painting would be easier.
>> Yeah, I was just thinking exactly that.
>> because it's just like
square footage almost.
>> It sounds so
much easier a problem to solve.
>> It's actually slightly more difficult,
because if an on-site
estimator comes to estimate your
painting areas with all this
furniture that's blocking
the views of the walls and, so on.
And also getting skill is
important in painting, so
we need to get to the square feet level.
So we looked at it, and
then we thought that
the moving is an easy
problem to solve usually.
>> Technically speaking.
>> Technically speaking.
>> I think, this is such a huge business,
if you guys can eventually just solve the
estimation problem, the pricing problem.
You've solved 90 percent of
the headache that all these home
service lead generation sites.
>> We agree.
[LAUGHTER]
>> So this is like a billion dollar,
this is huge.
This is the next Google potentially,
right?
>> Yeah.
Congrats.
You should apply to YC.
>> [LAUGHTER] Yeah.
Yeah, looking forward to that.
>> Okay. [LAUGHTER]. >> Great job guys.
>> Thank you.
>> Yeah, thank you.
>> It's good to meet you.
>> [APPLAUSE]
>> Hey.
Avichal.
Good to meet you.
>> Tell us what you do.
>> Hello.
Can you hear me?
Yeah.
So, we are Alma based.
So what we do is we help schools
engage in their communities.
So currently schools have problem tracking
their alumni once they graduate and
alumni experience is very broken.
Like the moment they graduate
they keep getting these giving
solicitation emails and nothing much.
And then so
with alumni base what happens is you can
create a private alumni network where your
alumni can sign up using the universal
log ins like Facebook and Linked In.
And then they can engage with each other
for mentorships, for job opportunities or
for signing up for
the reunions and such things.
And when those alumni are engaging,
the school would get insights like
how was this particular alumni
engaging with our Facebook page.
How many times he has like us or
how many times he has commented.
Also team sights so
they can prioritize their alumni.
And then they can reach out to the right
alumni for different opportunities.
>> Is this for colleges, high schools?
>> Yeah, K-12, independent schools,
and universities.
>> Do they all have the same problems
in terms of reaching out to alumni?
>> Yeah, universities definitely have
a different problem because they
are a different scale.
But schools have a different problem.
The biggest problem that the schools
have is they do not have updated alumni
data because the moment they graduate-
>> So,
it's just a basic, like all you
do is currently is basic CRM and
then in would be a hundred X
better than what they have now.
>> Kind of, but the way that they get the
data into the CRM is by the social media
integration, that's the key value that
we provide.So the moment you sign up,
we get your profile information from
Facebook, your education, your employment,
your current location.
So if someone has moved from San Francisco
to Boston and they updated on Facebook,
it automatically gets synched here.
>> Currently the way schools are tracking
it is by either putting up a Google form
on their alumni tabs on
the school website, or
by asking them to email or phone-a-thons.
>> So is their part of the problem
just connection to that person?
So once, like just emailing them, and
them actually seeing the email,
and then converting that person.
>> Yeah, but it's quite difficult,
like how many times you have responded
to any of your school emails.
And then once you get
the input from the email,
then you are going to need to
put in some in the database.
And you always need to manually update
the database next to a spreadsheet or
something like that.
>> Which type of customer
are you focused on right now?
High schools or college?
>> High schools,
>> High schools.
>> Interesting.
>> Yeah, [INAUDIBLE] ones.
>> From the from the end user perspective
why would they want to sign up?
Like they're having Snapchat and
Instagram, a Facebook,
a What'sApp, instant messenger,
a phone, an email address, a Slack.
>> Yes.
>> And this?
>> Yeah, I got it.
>> [LAUGH]
>> This is something a lot of people ask
me every day.
So but, I'm in startup school, like in
our group, I just posted to say, how
many of you want to join a new network of
your alumni, and why do you want to join?
Everyone was like, you we want to have
a private network because Linkedin and
Facebook are public networks.
>> Why doesn't a Facebook
group solve this?
>> Facebook group doesn't solve
this because you cannot actually
link to each other for job opportunities
or for mentorships, for one thing.
>> What do you mean?
Why not?
>> Sorry?
>> Why not?
You can just message someone if
you're in the same group.
>> No, but, I mean, how many times do
you actually get those notifications?
>> I turn off lot of my groups.
And then each of the group
is seven person these days,
because Facebook really want to
promote ads or something on the group.
So it's quite difficult in terms of
that engagement on Facebook groups.
>> But why wouldn't the same thing.
So, if I followed your logic,
you basically said you join a group,
you could, theoretically join a group.
>> Yeah.
>> Then people could reach out to you.
They can like your stuff.
They can post a bunch in the group.
>> Okay.
>> And
Facebook will notify you because
there's stuff happening in the group.
>> Yeah.
>> That starts to get annoying.
So you turn off the notifications.
>> Exactly.
>> Why wouldn't the same
thing happen for you?
>> Yeah, so here you can actually set
up different notifications when they
are signing up.
We asked them, how do you want to
contribute back to your community?
Is it by mentorship or is it by providing
job sharing opportunities, or by giving or
volunteering or planning a reunion?
So ask those, and based on that,
they can configure notifications,
and they would be getting
a digest based on those settings.
>> I see.
>> Yeah and so
you wouldn't be getting
all the common ones.
For example,
the class of 1990 is having a reunion and
you're on the Stanford group and
you're from the class of 2000.
Then you actually don't want to see that,
right?
I mean that's noise for you it's
not relevant for you in the group.
>> How many schools do you have currently?
>> To be precise, 94.
>> And how many per school,
alumni per school?
>> So, on average it's 5000 and it also
goes 15,000 depending upon the boarding
school or the kind of school it is.
>> 5000 alumni per school?
>> Per school.
>> Yeah.
>> Are on your site?
>> Yeah.
>> How many school?
>> 94, but-
>> 450,000 [INAUDIBLE].
>> No, we have our own like, million
records like some schools have more and
we also have signed up
some of the universities.
>> I see.
>> And of that,
like 25% have signed up on the platform.
Just so, was still,
you have to sign up on the platform.
>> I see.
>> And what are they doing
when they sign up to the platform?
>> So one thing is the admins,
like the online relations office,
they have different tools, where they can
post content and a weekly digest goes.
So the come and
they comment on this content or
they respond to any of the things,
like there's a section called needs.
If someone wanting help,
they can just post there.
So they comment, they respond to this.
This is actually that they
do at this point of time.
Are there any job postings?
They just come and apply if they're
looking for job opportunities.
>> So if I'm at the office that's,
who's actually buying this,
who's making the purchase?
>> The school is buying it.
>> Who's inside the school?
Is it the alumni association?
>> The alumni relations office.
Not the association, but
the alumni relations office.
Every school have a development
office in which the alumni
relations office is a part of.
>> Yeah.
>> So they buy it.
>> I see.
How much
are you pricing it for these people?
>> Yeah, we are pricing.
So we have three different plans, but
the pro plan is something which has lots
of our features, which is $5,000 per year.
>> Per year.
>> Yeah.
>> It would seem like one,
to get these users,
it would be right after they graduate.
>> Yeah.
>> Two-.
I feel like the missing feature here,
something along the lines
of school spirit.
I can't articulate right now.
I'm trying to think like what Avichal is,
trying to say is a lot of what you're
building is already on Facebook,
or these other things.
And people naturally create
these groups actually.
And maybe it's not the entire school,
but it's their friend set.
I'm in a group of 100 people,
my high school I think, so, but
what I, I don't know,
I'm not sure where I'm going to right now.
But I think the way to capture
these people, and the alumni folks,
the administrators can capture some
kind of school spirit thing where -.
>> Yeah, yeah.
>> Come back to the other side.
>> Exactly.
So
if you look at it the reason why you do
not want to give your contact information
is because the school is not providing
any value to you at this point of time.
So, but like if they could provide
some value hey like there are so
many mentors if you want any
mentorship and you just come here and
we would be helping you and
then there's a student aspect there.
For example,
the graduating class in the high schools,
especially when they're going for
their college applications or something.
In the counselling offices in the schools,
they can just match these people and
then help them get a better
counselling to get into college.
So this kind of small things that
they can do at this point of time,
which I think is not so structured, in
Facebook groups, or anything of that sort.
And, the key thing here is that only my
office can take control of all of this.
Like what is the activity
that is happening?
Which is not [INAUDIBLE].
.
>> What is success look like
for the alumni?
In a year,
they have to renew the contract?
And buy again how do they decide whether
to give you another five thousand dollars.
>> Sure.
>> I think it is about the number
of alumni that sign up and
then the kind of enrichment
that they get in the data.
For example, one of the schools so they
have only about six hundred records but
the total alumni stands is
like eight thousand people so
we also have an invite tool where
they can invite folks we can link in.
So now they have like 1,800
folks in the first six months.
>> Yeah.
So why do they want those records?
>> So again, for giving, or
knowing the preferences and
engaging them better and
providing them more value.
>> Sounds like there's
some downstream metrics.
Like, let's say I'm the alumni office.
>> Um-hm.
>> And I got all
8,000 records in the database.
>> Okay.
>> And the principal of that school?
Do I fire just the alumni guy who that,
that person is like out of a job now?
Cause they got all 8,000 records?
>> I think the place where
they're going to use it is, for
example, it's for mentoring.
Which is a huge value
that we are thinking.
But even the alumni officers not really,
they do not know what are the key metrics.
So we are also trying to help them
build those KPIs along the way,
when they're working there.
>> It sounds like the answer would be
similar along the lines of donations-
>> Yeah that's right-
>> Per alumni coming in, right?
>> Yeah that's the end goal,
if you look at a funnel-
>> So what are the tools,
then you have to build
them these tools right?
So you got all these people in, great.
>> Yeah.
>> So now how are you going to get these
people to donate?
>> To donate?
So we have a fundraising model where
you can put different campaigns,
like this is the annual campaign, this is
a campaign for a particular building, and
then you can make
donations on some already.
>> So, that's the critical thing, if the
alumni person is trying to figure out how
to get more donations ultimately, why do
all this extra work to try to get in.
Why not try and
go through their existing channels and
help them optimize what
they're already doing?
>> Sure, so what we have seen is when
they're just doing the station for
these donations,
actually the donation rate has come down.
For example in the schools, so
it It has actually come down,
when compared to the alumni, general
users just donating into nonprofits.
So these are just relating to cause,
and not to the schools, or
anything of that sort.
So what we believe, is like the premise,
or the hypothesis that we have,
is engaged alumni would
eventually turn out to be donors.
That's why all of this.
So you need to provide value before and
then ask them.
Don't just ask them.
>> Yeah, I see.
And is there data from, and I guess their
software's been around for a long time.
I have to imagine.
It's effectively CRM, software.
>> And the backend, just yeah.
>> Or I get emails all the time from
organizations that I've been involved with
that try to engage me, and have programs.
Is there data that says yes, if you engage
your alumni, they'll actually donate more?
>> Not yet,
I mean we have been for 18 months.
>> But none of it from other software?
Like there's mailing lists, right?
So are people who are responding to
mailing lists more likely to donate?
>> So, that's what the school
say at least, like we haven't,
I mean I ask them like,
who are your current donor?
So they say, yeah these are like
alumni who are engage who teaching, or
who come here for guest lectures, or
who regular respond to our emails.
But we do not know the exact numbers,
but that's what we get a sense of.
>> Interesting.
>> Is there a reason why you started
with high school versus colleges?
Because it seems like colleges
have much bigger budgets for this.
In fact,
they have entire offices for this.
>> Yeah.
>> And then,
the alumni had more money to donate.
>> [LAUGH]
>> Sure, sure.
So as with the schools,
surprisingly so as the schools.
But the reason why we initially
went behind colleges, but
then the sale seconds were really long.
There was a lot of additional making
process, and the procurement process, and
some bidding system, and some users.
>> But that's where all the money's at.
>> Yeah, we'll eventually go there.
>> We'll eventually go like once,
several times so we are just-
>> But I think the tools you build for
high school alumni administrations
are much different than the college ones,
way different.
>> Yeah, eventually.
Yes.
>> I think what you're suggesting is-.
>> Go to just the high schools?
>> Well, we've talked to you for
ten minutes, so
I wouldn't pivot the entire thing right
now, but I would really think hard about,
you know, the market opportunity
between high school and college.
>> No, I agree that the market
opportunity is huge with universities.
I mean, just the pure paperback math says
that, but the only thing is it's a lot
more sale cycle and it's just stagnating
the growth at this point in time.
But with the schools, there
are eventually two people in an office and
if you both agree to buy it, just buy it.
>> Yeah.
>> But
in a university, okay you
are the alumni relations director and
you are the second director and
you both agree.
Now you need to go to development director
and then you need to go VP of advancement,
institutional advancement, and
then the procurement office.
>> Yeah, but once you figure
out that sales price process,
I know it's difficult but once you figure
it out, it's probably similar in all
colleges, you just optimize that and
get really good at it.
Then people will start
coming in faster right?
It is a pain to get it, sort of.
>> Yeah.
What we are seeing is like, for example,
there are other players in this market.
We are just tracking of what
retails are closing the high rate.
They're selling the and
we are talking to a few people so
they say it's eventually how much ever,
I don't know they have really
cracked the process, but they have like
nine to twelve month sales cycles.
It's like a long relationship
building process.
They initially need to meet a champion,
then he, They echo for them.
It's really, really long.
Like a sales process.
>> Yeah, now I mean, that makes sense.
It makes a lot of sense.
I think the risk, really,
is that you end up.
The thing that you want for
highschool is actually very different
from what you want from colleges.
>> Sure.
>> So actually you have now
built a product for this market that
can't go to the more valuable market.
Like you end up in this niche that
is great because it was faster to
win that niche, but it doesn't translate
to the bigger more valuable niches.
Like the risk you run with that strategy.
>> True, but there is a slight overlap in
kind of the tools that we're building.
So, I feel that once we
get some insights here and
we also have a couple of universities, I
think 80 universities currently using us.
So based on those insights,
we are keeping a close eye on them.
And depending on that,
once we build some enough success stories
with them, like a case study, then we have
much better social group which might be
helpful in optimising the sales process.
>> When one I think this is what
Avichal was getting at earlier,
is-
>> Mm-hm.
>> These, the people at the colleges-.
>> Sure.
>> They already have a system in place.
Like a mailing lists, marketing,
campaigns and stuff like that.
Maybe you don't initially replace all of
it, but maybe talk to a bunch of these
people and maybe you can build
tools to help them do something.
I don't know what it is,
but something better.
And then start from there.
And it wouldn't cost as much.
But it gets you in the door.
>> Sure, sure.
Yeah.
>> Yeah, sorry but for
the other big thing,
everything you're saying makes
a lot of sense in the near term.
I mean if you look at, there's this
great story that's a very first banner
ad on the Internet had a 73% click through
rate or something insane just because
nobody had ever seen a banner
ad they were like what is that?
And so everybody clicked on it.
>> Okay. Yeah. >> And then obviously
people won't click on banner ads anymore.
>> [LAUGH]
>> I'm sure,
you really guys get notifications
that had crazy CTR and
we still actually surprisingly good CTR.
>> Okay.
>> And so
the risk you actually run is that by
trying to build your own distribution
channel that you end up having the same
thing happening to you a couple years from
now, when you're in a tough spot.
But actually, to me, it sounds like
you articulated it really well is that
the crux of what you want to do is help
these people raise more money, and
the channel is just the means
to helping them raise the money.
And another way to attack it is okay,
if you want to raise money.
What are the ten things
that I could build for
you that would just raise the money rather
them needing to channel to his money which
is just one tactic.
>> Sure so like a buy so
like a pilate, what I am doing is
volunteering with one of the customers
in the bay area school who is doing
the specific fund raising campaign so
we are building some tools for that.
To understand exact problem like what
are these tools that are required for?
If at all we need to just only sell the
fundraising aspect, do it for free, and
maybe take a cut.
So which makes the decision
making process also very easy.
So we're volunteering and
working and figuring out what is there,
what is there in fundraising.
>> Maybe put it a slightly different way.
It seems like there are already a lot of
channels out there between male, female.
>> Facebook, LinkedIn,
whatever, Messengers,
text, and building yet another channel
creates a lot of project risk for you.
>> Okay.
>> But actually,
just helping them raise the money
better through the channels that
they're already trying to manage would
be a different way to approach it.
And you might be able to have to actually
go through a much bigger market.
>> Sure.
Actually, I think that's the biggest
challenge that you are facing also.
So, initially, we built these tools for
the alumni relations office, where they
can, in the back, and they can do all
of this because they are the one who's
make the purchase at the end of the day.
So they need to immediately see
that instant gratificational there.
>> Yeah.
>> And improving their work.
But getting the alumni to sign up is a
very big challenge like you're mentioning.
So yeah, that's a big product
challenge that we have.
>> I can imagine literally if you just
made it easy, I go to some of the places,
I don't even, you click through
on the link and you're like, wow,
I'm trying to give you money and
you're making it so hard.
I can't do it on my phone, it's forms and
they don't fill in properly.
And then they validate
improperly from the JavaScript.
>> Yeah.
>> I mean, how, I'm trying
to give you money, why is this so hard?
>> Why is this so, yeah.
>> I can imagine literally you would just,
the alumni office would be so
happy if their conversion
went from like 5% to 12%,
they would just make more money.
>> Yeah, exactly, yeah,
that's what we are trying.
>> All right, cool, I think that's it.
>> Sure.
>> Thanks so much.
>> Thank you.
>> [APPLAUSE]
>> Good job
>> [APPLAUSE]
>> Can you answer one or two questions?
>> Yeah, sure.
>> So, questions.
>> Open Q and A.
Any questions?
>> [COUGH]
>> I have a question.
>> [LAUGH] Yeah, ask.
>> Yeah, so one of the challenges
that we face, for example.
So we're not able to
get close to our users.
So what is the generic advice
in terms of in that situation?
How do you actually get to
understand your users better?
>> Why are you not able to get your users?
You are here right now, right?
>> Yeah, but after a point of time, I need
again, tell back and then I can come.
So we are like switching,
our roles each of us come at
different times and then we stay and
then we kind of work with the schools, but
that has been ongoing challenge for us.
>> I think, where are you based?
>> Here in Mountain View.
>> In Mountain View.
And who are your users that are elsewhere.
>> So they're at different schools.
So the schools in Bay Area, and, yeah.
>> Sir.
>> They're from different states, sorry.
>> I see.
Okay.
Well, I think
the main thing is establishing
a relationship with these folks.
So, sitting down with them for
a little bit for, I don't know, a day or
a week and get the work flow
going is one way to do it.
Obviously, you can't
be everywhere at once.
I think you can build a lot of products.
You don't have to be right next to them.
>> I think there's two general
classes of ways to do this.
There's a qualitative and a quantitative.
On a qualitative,
there's all different scales.
You can literally have one
on one studies with users.
You can have a form
that you create online.
There's user feedback software that will
give you a bunch of qualitative feedback.
And if people upload something
that gives you some signal.
I've seen some companies use what
people would call a client council.
>> For you guys you take like a
representative sample of 12 to 15 people.
You bring them in once a quarter or
once every six months.
You show them new stuff.
They're NDA'd, they talk through it,
they give you qualitative feedback.
And it's like a very targeted way
to do this all in one sitting.
Then there's all the quantitative stuff,
so there's always metrics right?
Ultimately your KPIs.
And there's only really
four classes of metrics.
There's how many people use your stuff?
How deeply do they use your stuff?
How good is it at doing the thing
that you say it's supposed to do and
how much money do you make?
And all of your metrics
basically roll into those four.
And when your building features your
just mapping against those things and
you see whether those numbers move.
And so between those two, you should
be able to triangulate whether or
not a thing is working.
>> Thank you.
>> Do we have more time?
We're out of time.
>> That's it.
>> That's it?
Cool, thanks guys.
>> [APPLAUSE]