Ahmed Elsamadisi, CEO & Co-Founder at Narrator, talks about how they are building an innovative data platform based on his own innovation of data schema (called Activity Schema). We talk about how he’s come up with this brand new data modelling paradigm, how they’re trying to build a business around this now open-sourced innovation, their customer acquisition strategy & lot more.

The interview covers the following topics:

• The ways in which Narrator empowers data analysts in enterprises, offering quicker insights via the Activity Schema.
• Their journey to acquiring over 200 paying customers, including some prominent brands.
• Efforts to encourage data enthusiasts to amplify their innovation, fostering a ripple of positive online conversations.
• A glimpse into their current sales cycle.
• The intriguing "0 to 1" journey of Narrator.
• Insights into their team, funding trajectories, and a peek into their vision for the future.

how big are these typical deals?

Like, is it, I, I just wanna ballpark, and is it like a thousand

dollars, 10,000, or is it like a hundred thousand dollars a deal?

Like what, what does that look like on an average?

To be fair, uh, it ranges.

So we have companies like C b S and Levi, which are multi-billion dollar

companies using our product today.

We have companies like, um, Stapleton like, uh, code Academy, like

companies that are pretty large, but like not ginormous, fortune 500.

And we have companies like two people startups.

Hello everyone.

Welcome to the B two B SaaS podcast.

I'm your host ra.

Today we have Ahmed Elsi with us.

Uh, Ahmad here ran a company called narrator.ai.

Hey Ahmed, welcome to the show.

Thank you for having me.

I'm excited to be here.

Yeah.

Alright, ed.

So let's try to understand what your product does and

why customers pay you money.

Yeah, so narrator is an end-to-end data platform based on the

innovation known as the activity schema.

So you think about it as if you are going to answer, if a company's trying

to, their data analyst of companies are trying to answer questions.

Today they have to use many tools.

Each one of them is very expensive.

From modeling data to understanding how it flows, to using it in bi, to

ensuring its quality to you doing analysis, to sending it to a third party

product, and so on, and so on and so on.

Including debugging and investigating and exploring and all these

pieces and with the, uh, modern day stack that would require.

About maybe six or seven different tools and often multiple

different roles of the company.

A data engineer, a bi engineer, data analyst, data scientist.

And what we do at Narrator is we have, um, built one single platform that is,

um, able to allow someone to answer questions from end-to-end in 10 minutes.

So is it like, you know, feed your platform takes in a bunch

of data that a, a typical enterprise has, or a typical company has, and

then you can just query, query it?

Is, is it something like that?

Similar, I think, um, we push for, uh, we live on

top of your data warehouse, so we have access to all your data.

And one of the things that I mentioned is that it's on top of an activity schema,

but that means that we help you model using this innovative modeling paradigm.

It makes modeling your data a lot simpler.

So instead of 700, 800 tables, you only have one table.

So it's a single table approach to all your data.

So you have one table, and because we have that table as standardized,

it's easy for our application and our product to really work with

you to help you do a lot of things.

So it's not like you have a free sequel, like write whatever you

want and do all the work yourself.

We're really guiding you throughout that process, which is how we're able to get.

The level of quality and speed in answering questions.

So I'll come to your product in a while.

So I just wanna understand who you're primarily selling it to, right?

Like, how does your customer base look like?

Yeah.

So people often ask, what industry are you part of?

And I'm like, not really an industry.

Our ideal customer profile is really unique.

It's usually, uh, one of two companies.

Either you're small, You have one or two analysts and you wanna set up a

whole data stack and you can do it with narrator very cheaply and easily, or

you're really large and you already spent a lot of time setting up data.

You got angry at data, you fired half your team, and now you're like, oh shit,

I need to answer the same questions I used to do with half the resources.

So usually it's people who are really cost sensitive and want

to do a lot more with data.

We often, yeah.

So we.

Usually it's like a company that cares a lot more about the output of

data versus the um, system of data.

And that's why kind of we see a lot of our comp customers are in Europe.

'cause Europe tends to be very high security, very sensitive with data.

So they want highest quality and.

They care a lot about like bang versus buck.

They're like, if I'm spending much money, what's the value and return of my money?

And narrator's return is really, really great.

And it's designed for a minimal user experience where like you don't, you

have people answering questions, but you don't have like a thousand people

All right, so,

or actuality

let me sort, let me try to understand this a bit more, uh, clear.

Right.

So how many sort of, uh, customers do you have on your platform as of today?

Like I'm talking about paying customers.

Yeah.

Paying customers probably a couple hundred.

A hundred.

And how big are these typical deals?

Like, is it, I, I just wanna ballpark, and is it like a thousand

dollars, 10,000, or is it like a hundred thousand dollars a deal?

Like what, what does that look like on an average?

To be fair, uh, it ranges.

So we have companies like C b S and Levi, which are multi-billion dollar

companies using our product today.

We have companies like, um, Stapleton like, uh, code Academy, like

companies that are pretty large, but like not ginormous, fortune 500.

And we have companies like two people startups.

So it really ranges from the kind of companies we talk to and

different people value different

what's your sweet spot like that you typically, you know, aim for,

like typically your sales teams go after, like, what's, what does that look like?

yeah, we really focus on.

Companies with a high ratio of data analysts to data engineers where there

is a lot of people who want to use data, which is often data analyst's

job, and not a lot of people building and maintaining data infrastructure.

So that's kind of the key piece for us,

So I'm still looking for a number.

Right?

So, so just, just trying to understand, right, is it still a thousand dollars

deal or is it a $10,000 deal, or it's a hundred thousand dollars deal Because

all my subsequent, all of our subsequent interviews are gonna based on that, right?

Because your sales team or everything is going to be very, very different.

Yeah, yeah.

Right now we're really focused on $10,000 deals around time.

Got it.

That, that makes a lot of sense.

Right.

So, yeah.

And, and just like, uh, so let's, let's try to understand

your top of funnel, right?

So like where, where you're getting all of these customers from, like

what's really been working for you?

How, how are you sort of able to drive all of these customers to your platform?

Um, so far it's been really two pieces.

One has been word of mouth.

Content, content.

So we focus a lot on sharing the ideas that we have.

Um, we have still never done paid marketing.

We've done $0 in paid marketing.

Um, we don't like host events or anything like that.

All we do is share the ideas.

I think, um, in our open source community, the people who follow

the activities schema, there's a lot of strong believers there.

Recently somebody built an entire course on it.

So like people really appreciate this open source, um, approach, and then they

come in and they end the binary narrator.

Got it.

So just before getting into this, right, I just wanna get a sense

of like, how does your growth look like over the past 12 months?

Right?

So how many new customers did you manage to acquire over the past 12 months?

I think we doubled, I think we nearly doubled our size.

got it.

So that makes a lot of sense.

I'll just talk about, like, I just wanna understand a bit about,

you know, word of mouth, right?

I mean, word of mouth is such a catchall phrase.

I mean, you could, you could just say everything, right?

How exactly are you controlling this particular element?

Like what's really happening?

Where exactly are people talking about you and how are you sort of driving it?

Yeah, so when I say word of mouth, I really mean content, so I

really mean like really heavy on content.

Um,

And what sort of, what type of content are you talking about?

Like say, can you just

we have,

that?

Yeah.

yeah, we have, we have several layers of content.

So we have one which is our blogs.

So we do a high detailed blog.

We don't write blogs that are like 10 things.

You should really.

Click on, like we do like thoughtful blogs that are trying to share

philosophies, seconding is we share, um, activity schema work.

So activity schema is the foundation of what we do.

It is the approach that allows us to answer questions fast and that has like

meetups and follows and people write about it and people talk about it.

And there's a huge amount of things that people talking

about this modeling paradigm.

Really only two modeling paradigms today.

There's a standard way of doing data modeling, which is known as star schema.

And then there's this thing called activity schema.

And it's been really impressive for people when they start setting

up a system to say, oh, what?

How do I model data?

And there's two options now instead of one.

And that is what gets people really interested in narrator.

They're like, wait, wait, wait, wait.

Oh, you guys created this modeling paradigm.

Okay, I wanna know more about it.

So that's usually what we focus on is that content.

Like, how, so?

I think I, I'm not, I'm not very clear about this particular schema.

Right.

Even though I'm an engineer, so I can't really deep dive into it.

But when you say you created it, I mean like, and like how

exactly did that turn out?

Like what was the history here?

I.

How exactly are you creating this, you know, open source community?

I mean, getting people to talk about something like that is

not an easy thing as well.

Right?

So how is that even working for you?

What are you doing on a regular basis?

Just talk a bit about that.

Right.

That sounds very interesting to me.

yeah.

So, um, The, how did I create it is a good question.

Is, is it, is, did you created, by the way, is that, is that it?

Oh, okay.

Yeah.

So I spent, um, several years trying to, um, solve a problem in data,

which is if you go really, really, really low and you're an engineer,

so in core, core data, you have to join data based on foreign keys.

That's like the essence of data.

When you deal with like large scale data or data warehouses,

those foreign keys never exist.

Mm-hmm.

So how do you build a system that can connect

data without foreign keys?

Mm-hmm.

It's a really hard problem.

It's like unknown.

So that's what the activity schema was designed to solve.

It's an it.

It's by using standardized, standardizing by standardizing the data model and

relating data using temporal joins, which is the idea of using time and

customer as the only way to still create role level joins is the innovation that

we spend time building and setting up.

And how do you get that working?

So people who, uh, do data, who do behavioral analytics, who do product

analytics, who, um, Started picking it up because the D the community of open

source data modeling and data engineering thanks to D B T grew really quickly.

And people started talking about like, okay, 'cause data's relatively new, like

you're young, the data engineer industry.

So people were like, okay, how do we model it?

And I think we started going after, we used to call jaded data people, people

who've been, who've built more than one data system who know why it doesn't work.

And they can just like, and we would say whiteboard them to death.

We just talk to 'em, be like, look, look, look, here's what happened with me here.

I did this six times.

And that started getting those couple of people who were jaded to start talking

about like, this shit doesn't work.

And slowly and slowly, a lot of people who were doing like data content, uh, or

creating product analytics were started talking about like, oh, This idea.

'cause we are a time series table at its core.

So people say what happened is that it kind of caught on to this other

world, immersed into one where people were like, oh, I have event data

so I must be using activity schema.

And I'm like, I.

Kind of, and then we kinda were like, okay, let's, let's run with it.

Let's be like, yes, you can do it, but here's the rules that

makes it an activity schema.

So a lot of people heard about it and were like, okay, it's just event data.

And then we got a lot of press and power from the event data community.

A lot of people were like, oh, actually it's more than that.

It's such a data modeling paradigm.

So we got a lot of people who were like, Ooh, I want like, I love this.

And then we got like the our core initial group, which started it all, which

were people who were like, this, the standard approach doesn't fucking work.

I wanna use something different.

And what we've been doing with this community, It's continuing to foster it.

There is open source, um, uh, uh, open AI packages.

There's open source D B T packages.

There's open source, uh, content.

There's so much information on how to work.

We have what a spec to explain to people how you actually use it.

And we continue to put more and more content on what and how you answer

questions using this approach.

And between like the Slack channel and the online people just

started talking and it clicked.

And I think when you have an idea that.

Clicks with people and they're like, oh, I get why this is so different.

And I get why it's so important.

It just kind of naturally you grow with it.

And then we recently dropped our price really?

Uh, we a lot by over like 90% so that we can actually let

people come in and try narrator.

'cause we used to have a, we have a lot more people who love the activity schema.

It was like, Of people who are like loyal to it versus the couple

hundred that are using narrator.

And we wanna do is we wanna shift that.

So by dropping our price from being like a hundred to 50 deals down to

10 deals, its come in a lot cheaper and then grow into our bigger deals.

Got it.

So essentially what I'm understanding is like you've got this brilliant idea

of, you know, how you sort of model data and then you started creating, enforcing

this online community around that, and then just, just sort of building

a com company around that as well.

That's, that's how it's working for you so far.

Uh, yes.

The timeline's a little off because we actually built the

company and started a consulting company far before we released it.

But what we found at one point, one of our co-founders was like, Hey, I

think that people will follow this idea that we're using as our secret sauce.

And I was like, well, he is like, we should really.

Open up our secret sauce and start talking about it and letting the

world know that there's a secret sauce that makes us able to move so fast.

And I was like, that's our competitive advantage.

You can't do that.

And he was like, no, I think it really might inspire a lot of people.

So we ended up opening

got it.

So, so I mean, that, that actually sounds like a wonderful journey, right?

So just want to get a bit more specifics here, right?

So, so what do you do on a regular basis to foster this community?

Right?

I mean, I see that you've open sourced it, but like how do you

get people talking about it?

I know people are very interested in this paradigm and all of it, right?

But I.

What, what, what is it that you exactly doing?

You talked a bit about Slack channels, you know, doing things like those like,

but what's really happening there?

Like, where do you spend all of this time and, you know, what, what,

what are those initiatives to be?

Very,

Yeah, so honestly, We, we spend a lot of time with haters

and people who are challenging it.

So that's kind of our main focus.

Like for people who are like, oh, this is interesting.

Anyone who's interested online about it, I try to meet with them.

I try to talk to them.

I try to give them examples.

I try to show them how it works with narrator.

I try to really, really dive into anyone who believes in it.

So we're not like mass, we are, we do our blogs, we do write

content, we do share stuff.

We're not mass.

Marketing our open source work.

Yeah, we're planning on doing it later in our time, but right now

we are working on our, what we call our early adopters, and we wanna

trying to understand like, where is, where are these

conversations happening, right?

So what's that

Oh, LinkedIn

happening?

It's, it's on LinkedIn.

Okay.

yeah.

Usually LinkedIn is where people are like, this is stupid, this will never work.

And then I'm like, can you gimme one question I can't answer?

And they're like, what about this?

And I'm like, here's how you would answer this question.

Here's

to be honest, I'm a bit surprised, right?

How, how do you even, and then like, just give me some numbers here, right?

So how many people are we talking about?

Is it a hundred people?

Is it a thousand people per typically, you know, engaging in those conversations.

Engaging in our, like, we would probably get, like,

some of our content gets like between like in the, like some content gets

like 20,000 likes, 30,000 likes.

So you're talking about like maybe 50 to a hundred thousand impressions.

Uh, in our big things you'll see a couple hundred comments,

so it's not like a humongous.

But again, these are the, the people that we are talking to and the people

who care are the people who are core leaders in the data community.

They're not like a person who doesn't, who doesn't know.

It's people who are passionate about data modeling.

That's people that we've signed.

We focus on a very, very, very, very, very subset of people who are making decisions.

Eventually, we're gonna open up more and more and more, um, to like, People

and talk about like why a single analyst and narrator is all you need to answer

questions and why you can unlock the product activity schema and like, do

you have more than a hundred models?

That seems like a lot.

Product activity schema, like there's gonna be a lot of pushing people to,

uh, give it, but right now it's, we need to become fully self-serve because

narrative is still not fully self-serve.

It's like now like 80, 90% self-serve.

We need people to be, um, aware and the community to be sharing content,

not just on activities schema, but also on narrator, so that we

can go out there and be like, you know, share and get more attention.

Got it.

So help me complete the funnel here, right?

So I now get a sense of where you're getting all of these leads from,

from all of these conversations.

You know, people out there, right?

So what happens, how do you convert them into a paying customer, right?

So what's happening?

Like what's, what's that journey look like for you today?

I.

So journey today is very, very simple.

We have one call, so they book a meeting online.

We have one call and we walk them through our product and we sign a one

year contract with a one month optout.

Okay.

That's it.

So we say, look, I want you to be committed to using our product.

It is a learning curve.

You're gonna spend time learning it.

I'm gonna show it to you.

I'll sign a contract, but you can opt out after a month so you don't have

to like, it's kinda like a trial, but with the contract signed before, uh,

so you don't have to, so we don't have to like, convince them to, uh, convert.

We get the contract before we start using the product so that we have no, like

six months POCs and stuff like that.

you mentioned it's not a self-serve product, right?

I'm assuming you need to do some sort of integrations into the, you know,

data warehouse or whatever that is.

No, no, that part is Selfer.

When I say, I said it's 80%, so I mean like you can't go

online and get an account

Mm-hmm.

by yourself.

You need to talk to someone.

So I like, we still talk to you and ensure that you know what you're getting into

and we make sure that you sign and we make sure that you're committed and then

you get started with the product and then

I am surprised you're managed to, you're managing

to close it within just one call.

Like what's, what's, what's like really working

Well, the price is low right now.

Well, it's 500 bucks to get started, right?

So it's like you're getting 500 bucks a month to get started.

These tools, on average, we support, like, we're like a nine different pieces.

Each one of these pieces, if you go find the cheapest competitor, you're

gonna pay a couple thousand bucks.

So just like data modeling, bi, all these things cost a lot of money

and having one place for all of it.

So even just on price, we are pretty like low in the, in the, in the, um,

In the, uh, landscape of data tools.

That's kinda what we're trying to get them into.

Um, and then usually our demos, we try to like be really aligned.

I'm like, what question are you trying to answer right now?

How long will it take you to answer it?

They say, like, a couple weeks.

I say, great, I'll do it live in front of you with our tool, and then

if you sign up and I'll, I'll send you an email exactly how to do it.

On your own data and people often do that.

They get started.

So, and then, so yes.

Right now our conversion rate to buying is pretty high.

It is the, I think it's like once you actually get started connect to

warehouse, it's like 73% actually end up buying and continuing.

Um, but there is a huge drop off of course, when between the first

call to deciding to move forward.

That's, but we're okay with that because what we want is the people

who are using our product to be loyal.

Happy customers.

And that's kind of what we get right now is our users are like

really happy, really loyal.

That's how we've

how do you plan on expanding this?

Like have you seen any sort of expansion stories so far?

Like, I mean a $500 deal getting converted to let's certain thousand dollars deal?

Do we have

Yeah.

who are actively trying to push for that or,

that's, yeah, that's most, um, data tools right now.

That's how they price, price it.

So like if you buy like a looker, Average contract size is like around nearly 200,

250 K, but that's because like users go up, data size go up, like people do more.

So I think that for us, like we have two, we have two uh, growth avenues.

One is users.

'cause you only get five users when you start and people always add more users

and they're a hundred dollars user.

So you add a new team and new team and new team and new team slowly grow.

The second thing we do is we, because narrator has templates,

We are able to answer.

If you're like an eCom company or a sa, a SaaS company, you can ask narrator

a very in-depth question and our internal team will answer it for you.

So like, it's kinda like consulting, but instead of being open-ended, it's like we

only, we charge you a flat rate of $500.

Got it.

um, that one grows like most of our larger companies, like

ask like five to 10 questions a month.

So just like, That number really ends up being the deal, becoming

like now it's $5,000 and just questions answered a month.

Plus you get your platform and, and the workforce to answer The more

users you have, you get another five to $10,000 a month from there and it

ends up being a really nice contract where people are getting lots of

value using the product and growing.

Right.

So you mentioned something called.

Consulting.

Right.

So I mean, I'm still assuming your platform is answering

all of those patients.

For them, it's just that you're helping them sort of do that.

Yeah, but the benefit of narrators, because the

data model is standardized, every question we answer can be reused.

So that's the magic of narrators.

So when we have, we have like a couple hundred templates that can answer like

most e-comm questions, like it's really rare that a new question comes in.

So instead of when a new question comes in, if it's, if we have

a template for it, we just use the template and and deliver it.

There's no template.

Then we built it and our product's really good at make

building answers questions fast.

So we built analysis and answer it, and then we added to our template library.

So that's how we were able to kind of scale this like consulting feeling,

experience with, um, mostly tech enabled.

you charge for creating those templates?

Is that what you're saying?

For the first time when somebody.

Um, no, we, we charge for, for the question getting answered.

So whether we have to create a template or don't create a

template, it doesn't matter to you.

Don't you just get, you get, you ask your question, you get the analysis.

Alright, so in the light of time, I wanna move, move forward, right?

So just, just sort of wanna understand it.

How did it all start for you?

Right?

What's, what, what, what's the backstory there?

When did you start the

Yeah.

So

how did it all, you know, started?

I.

yeah, the story, the origin story is really simple.

It's, um, I used to run data for WeWork.

So big company growing really fast, um, built their data system once,

twice, three times, four times.

I kept refactoring it and I kept, I, we had a, like a little over a million

dollar budget on tools and we still couldn't answer questions fast enough.

It was really bad.

So our c e o asked me to go like, find examples of what companies

other companies are doing.

So I talked to Airbnb, I talked to Netflix companies and

everybody was doing the same shit.

Which is what I was doing too.

You're spending a lot of money and putting people out.

The problem, and I, I had a hypothesis of what was causing the need for all

these tools and people, and my hypothesis was that because joins don't exist,

you have to hack around it and the only way to hack around it is thousand lines

SQL queries and nothing is good when you have thousands of thousand lines

eagle queries on top of each other.

So I left and with the idea that we, if we can.

Create a way where you can answer questions where foreign keys don't exist,

that is not a hack that is more standard and structured, then you can get rid

of all these dependencies and layers.

And the activity schema came out of that approach.

Essentially, time becomes your foreign key.

Right?

We just basically joining all of your tables.

pretty much, yeah.

Time.

Like, uh, more like nuanced time, but yeah, time and

customer become your foreign key.

But standardizing in a way where you can actually answer, you need enough

ways to connect with time so that you can do like, 'cause the problem

with time joins is that you can cascade and blow up and get 20 rows.

So you wanna make sure that you guarantee you can only

pick one row using time joins.

And that's really hard because you, you want to join data without

ever duplicating or dropping rows.

So you need to be very careful of being able to join stuff where

time doesn't exist or where time exists, but, and it is duplicated,

but you still have to pick one.

So that's the innovation that took us like, I think it was three years.

it's been three years, is it?

It's 20, 27 years.

Okay, so

Yeah, but three years to get, yeah, we started, um, early 2017 and,

it took us, the, the biggest thing was, um, getting that thing to actually work.

So we took, spent three years to try to even get a P O C of this

like way of connecting data.

Um, and then it took us another two years.

Like to get a product that we're proud of, like, you know, it takes time to build and

and you've been closing all of these deals in the past,

you know, one or two years, right?

Primarily trying to sort of build a

Yeah.

So most of our deals, yes.

Last, last two years have been like the, really the big, uh, growth spike.

And how big is your team?

As of today?

So we're a pretty small team.

Uh, we just downsized, so we're like a little under 10.

Um, so keeping a really small, we used to have a larger sales team, but we recently.

When we dropped our price, we got rid of our sales org.

Um, mainly because like salespeople, you can't have a, you can't afford to

have a salesperson if you're selling a deal that starts at 500 bucks.

So when we were selling them at a hundred thousand, uh, price point we're like great

sales org and people, and SDRs and AEs.

But if we're dropping our price with this idea that we can get actually

let more people use our product, that having a huge price barrier, uh, we

decided to kind of sif away from sales

Yeah.

And then have you raised any external funding so far to build the company?

I'm assuming you are

Oh yeah, many rounds.

like how much in total?

Just to.

I think around like 13 million.

13.

Got it.

And then what, what's the vision here?

Right?

So what, what are you planning to do with this company?

I mean, where's it gonna go in the next four to five years?

Yeah, so four, five years.

Um, the world, I think my goal is to allow anyone to answer

any question by an expert.

So right now we see you're able to answer questions with narrator.

Yourself and you can do everything you need with a narrator.

Um, soon you're gonna be able to do more and more with a narrator.

Um, and we're gonna get to a virtual analyst.

So we are training our data based on, because our data standardized,

it makes it really nice for LLMs.

So we're gonna allow you to actually train your data based on your answers on how to

answer your questions, which, because of the activity schema, there's a great blog

post on it makes it really game changing.

League better than normal.

Trying to do LMS on top of tables.

So then the next path for that is why are, you know how we're

building templates 1,500 bucks and we're answering questions with it.

If you're an analyst in a company and you built something, why don't you share it?

So we wanna be the first ever like shop for well thought out data analysis.

So that's where we're going.

So right now, instead of like, you can sell it for free, you

can open source it, you can do it, but instead of selling like.

Products or selling, uh, things or selling idea like, or

writing blogs on your analysis.

You can actually give people a well thought out analysis.

And I think that's gonna be like, that has never been done in history of data

where you can actually run an analysis from someone who did it using a di using.

On your company that was built on a different company, like with One

click, and I think because activities came standardized, we're able to

share across companies really easily.

We've been doing it for now two years and it's been working really well for us

and we wanna expand that for everyone.

So in my dream, like as part of dropping the price is well wanting more people

to come into use narrator, and then we would eventually just mostly be an app

store where any question you have, you can find an expert who's done an has.

IT app will tell you what they did with it.

It has thoughtfulness and you can run it with one click, and now

you're, the whole world is sharing and making better decisions.

Yeah, that makes a lot of sense.

Yeah.

So all Ed, thanks for taking the time to talk to me.

Hope you scale narrat to much, much greater heights.

Thank you.

I'm looking forward to.