I was early Marketing at Uber spending 5 years building Brand, Performance, and Lifecycle Marketing Data Science teams. I got to work with & learn from world-class marketing, data science, and martech leaders. Over the course of my career there I've measured over $1Bn in Marketing Spend and estimate that I've helped drive $200M+ in incremental revenue and cost savings . 

Then I moved on to a few early stage startups and scaleups (both as a FTE and Consultant) where I learned how to be even scrappier when building Marketing. 

Now I help Marketers make smarter decisions through my newsletter & podcast on Consumer Tech Marketing with how-to guides, case studies, and stories. 

I'm doing this because I love the marketing stories and people behind it in a world where Product seems to get most of the credit. There's so much magic in Marketing and I want to share more about it. 

Check it out here → https://experimental.beehiiv.com/

Outside of my newsletter and podcast, I love spending time with my 2 beautiful kids and my incredible wife in Amsterdam. We used to love traveling and eating out but you know... kids ...  so now I love cooking and exploring all the fun things little kids can do!

Sundar Swaminathan

One of the hardest things about having kids has been reconciling my professional identity before vs after. 

Before, I cared so much about my career. 

Being raised Indian and American, what I did defined who I was.

Even as a kid, you’re asked “what do you want to be when you grow up?” 

“Doctor. President. Astronaut”

What you wanted to BE was what you DID. 

Early career, I’d chase promotions and titles and cared about so many small things. 

Post kids, I couldn’t give a sh*t. 

And I don’t mean I want to do sh*t work. 

I still have pride in my work. 

I mean I don’t care whether I get that next promotion or whether I am “changing the world” or working on something cool. 

I don’t have career ambitions. 

I just want to hang out with my kids. 

And I am working to live not living to work. 

But the gap between how much I care about my kids vs my career is widening. 

So much so that I now have an aversion to the idea of going back to a job. 

Like actually disgusted. 

Lol my friends have described this as me just having a “mid life crisis” but I think it’s something deeper.

It’s an existential one. 

One where everything I had been working for 15 years towards now carries zero meaning and significance in my current view of life.

Former Head of Marketing Science at Uber where I optimized $1Bn+ in spend across Brand, Performance, and Lifecycle → ✏️ & 🎙️ on Growth and Marketing in Consumer Tech

No alternative text description for this image

I’ve been using A.I. for over 30 years.

I use it for literally everything:

- Writing this post 
- Finding new recipes
- SQL when I was an FTE
- Speeches when I present 

The list goes on and on. 

But the weird thing is every time I tell someone I’ve been using A.I. that long they question me. 

They call me a liar.

That I’m just saying it for the likes. 

It hurts. 

But I can’t do anything to change. 

I was born with Average Intelligence. 

I was born with A.I. 

It’s who I am. 

Note: this post was written with A.I.

Former Head of Brand Science at Uber | Optimized $1Bn+ of Marketing | Author and host of experiMENTAL ✏️ & 🎙️ on Marketing Strategy and Science

CEO: Simple question.Should we do brand marketing?

CMO: Simple answer. Will you give me 6 months without pulling the plug?

CEO: That's a long time.

CMO: That's likely how long it takes to see brand impact compound. If you'll panic before then I'm not going to even bother. 

CEO: What happens at month 6?

CMO: Your performance CAC will do one of two things:
1. We either keep the same volume and lower CAC 
OR
2. We keep the same CAC and do higher volume. 

Depends on what our strategy and goals are.

CEO: How do we know it's working before? 

CMO: Proxy metrics, share of search starting to go up, increased "direct" traffic.

CEO: Those sound fluffy...

CMO: This is exactly why I asked for 6 months of you backing off, but here's another question for you that might change your mind. 

Are you happy with the current performance?

CEO: HECK NO! 1. We're burning profit 2. Our CACs are skyrocketing 3. The quality of users is really going down. 

CMO: And why do you think that is?

CEO: Idk.

CMO: I do. We ran out of people that know about our product after we picked all the low hanging fruit. 

CEO: Ugh that sounds right. We've been in market for a few years.

CMO: Exactly, but our product is awesome and we love it and we need more people to know about it.

CEO: Hmmm that makes sense to me.

CMO: Here's the best part... our competitors aren't doing this which means we can get a head start and start building a moat that both improves our growth AND slows theres down.

CEO: Ohhhhhh siiiiiick! Okay, you're right! Let's f*cking go! 

CMO: Heck yes... just remember...6 months. 

--
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this post later

CMO: We need to invest in marketing data science. 

CEO: Why now? What changed?

CMO: We did. At Series A, we had 1 channel and 1 market. Now we have 3 channels across 10 countries.

CEO: So just add more rows to the dashboard.

CMO: Not good enough. The measurement approach that got us here breaks at our current scale.

CEO: What do you mean breaks?

CMO: At Series A, I could eyeball attribution and make calls. Now, channels are cannibalizing each other, and I can't tell which ones are actually driving incremental growth.

CEO: Give me an example.

CMO: We're spending in paid social and paid search. Both show great ROAS. But when we scale one up, the other's conversions drop. They're fighting over the same customers.

CEO: So what do we need?

CMO: Incrementality testing. Holdouts. Geo experiments. The stuff that actually measures causation, not just correlation. But, I can't do it myself because I don't know how to. We need to bring in data science.

CEO: That sounds expensive.

CMO: Way cheaper than wasting budget on channels that aren't working. At our scale, a 10% measurement error is a seven-figure mistake.

CEO: Why didn't we need this before?

CMO: At Series A, our biggest risk was moving too slow. At Series C, it's moving confidently in the wrong direction.

CEO: How do justify upgrading our measurement?

CMO: The cost of being wrong is bigger than the cost of certainty. We're past that point.

CEO: Alright. What do you need from me?

CMO: Budget for proper testing infrastructure including data engineering and patience for answers that take weeks not hours.

CEO: Done. Now why are you texting me at 2AM?
 
CMO: Because wasted marketing spend never rests. 

--
🔔 Follow @Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this post later

CEO: I need 30% more keeping CAC flat by Q3. 

CMO: Sorry, that would break physics. 

CEO: What do you mean?

CMO: Between high volume, fast speed, and low cost, you can only pick two. 

CEO: Explain more.

CMO: Sure thing. If we want:

High volume + fast speed → High cost

The only way to pump up volume quickly is inject more spend but it'll be inefficient and so cost will spike.

High volume + low cost → Slow speed

To get high volume at low costs we'll need to do everything from better campaign structures, experimentation, data pipelines, and even upper funnel campaigns. That'll take time.

Fast speed + low cost → Low volume

If we need to get to lower costs quickly, the easiest way to do that is to cut spend which means low volume. 

CEO: So there's no way to get high volume, fast speed, and low cost?!

CMO: No. My hands are tied. It's literally Marketing physics. Anyone, including your other CEO friends, that says they can do this is lying or hiding something. 

CEO: Crap. Are there other laws of Marketing physics?

CMO: Yes. At least four more. You should read this article by Sundar. He writes about Marketing Strategy and Science from his time at Uber.

CEO: Sounds great. Send it over.

CMO: Here you go → https://lnkd.in/gG3UkXnm

Jr Marketer: Hey, I'd love to become more data driven 
Manager: *slacks 10 articles below*
Jr Marketer: *1 day later...* Whoa. 🤯. IQ +1000

--

1. Everyone tells Marketers to be more data-driven, but not everyone knows where to start. Here are 10 of the most common analytics terms that's a good starting place.
https://lnkd.in/gZA9-wce

2. We need to get smarter about measuring marketing and here are 5 different tactics we can use to measure what we're doing.
https://lnkd.in/gw3mXHGh

3. MAU is a KPI for us but we've been measuring it all wrong. This is one of my favorites on hwo to properly calculate MAU.
https://lnkd.in/g9MEku7h

4. As we transition to increasing personalization and need smarter segmentation, I really reading how Uber did it with their 600M+ customers.
https://lnkd.in/gHXiYMSg

5. We're coming up in 2026 planning and this one helped me shape how we're going to be doing planning and target setting for our team. 
https://lnkd.in/gmH8VV8T

6. Given we don't always have analytics resources at our disposal this one is great for when we need to run a pre / post analysis ourselves. It's not about perfection but just slightly smarter decisions.
https://lnkd.in/gbbw4yz7

7. We know how important LTV is to everything we're doing but turns out we had been doing it all wrong. I took this article to our Finance team and we've rebuilt it.
https://lnkd.in/gxUqcxcw

8. This one is just WHOA! This article is why we've added a measurement strategy to our campaign briefs and why our data science team loves working with us again.
https://lnkd.in/gPrr-tAF

9. This is a complement to the LTV article above but it's how we're defining CAC moving forward. It helped me align with Finance and why we're now BFFS.
https://lnkd.in/gmvsqxvF

10. This is just a fun article on how Duolingo turned around slow growth and grew 900% by being more data driven
https://lnkd.in/gWDnDj9X

Former Head of Brand Science at Uber | Author and host of experiMENTAL ✏️ & 🎙️ on Growth and Marketing in Consumer Tech

CFO: Our attribution model is amazing!

New CMO: Nice! I won’t be using it much. 

CFO: What? This cost us a ton of $$

CMO: It tells us what happened, not what we caused.

CFO: What’s the difference?

CMO: Attribution says Facebook drove 1,000 conversions last month. It doesn’t tell us how many of those 1,000 would’ve converted anyway without Facebook.

CFO: Wait. You’re saying the model is wrong?

CMO: I’m saying it’s precise and useless. It gives us confidence in the wrong answer.

CFO: How is that possible?

CMO: Because customers see ads everywhere. Attribution gives credit. Incrementality measures impact. Facebook might get last-click credit, but the customer already decided to buy after seeing our TV spot.

CFO: So we’ve been optimizing for the wrong thing?

CMO: We’ve been optimizing for who gets credit, not who deserves it.

CFO: Then what should we be using?

CMO: Holdout tests. Geo experiments. Incrementality studies. They’re messier, take longer, but actually tell us what’s working.

CFO: Why didn’t we start with that?

CMO: I don’t know but I just got here. Most of the time we confuse precision for confidence.

CFO: I’d rather be roughly right than precisely wrong.

CMO: That’s the mindset shift. We’re gonna fix this! 
–
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this later

Who reads my newsletter on Marketing in Consumer tech? 

Marketers in consumer tech. 
Also data scientist.
And founders. 

When I set out to write this newsletter, the biggest doubt I had (which is most people’s fear) is that no one would read it.

I’d just be writing into a void and while it gave me joy, it wouldn’t be a business. 

Almost 50 articles later and a year into it, I know that’s no longer true. 

50+ articles. 
2,800 readers.
60% open rate. 

And most importantly?! 

The audience I’m writing for is who reads it.

Marketing leaders across the funnel from Brand to lifecycle at companies like Uber, DoorDash, Apple, AirBNB, Spotify, Google and so many more. 

It’s just motivating to be able to take a step back and see the daily efforts translate to something. 

Anyway, Im excited and think my best work is yet to come. 

Thank you to everyone that supports me! 

Onwards and upwards.

CMO: We need to A/B test this launch!

Me: We can't. But here's what we can do...

Propensity matching.

CMO: Wait, why can't we A/B test?

Me: I don't think we want to hide the launch of our loyalty program from 50% of our users. 

CMO: Fair point. But, I'm being asked to show the ROI of the loyalty program. Will we be able to do that?

Me: Yes! We can use propensity matching.

CMO: Say more fam.

Me: So in a normal A/B test, we randomize users before and then measure the difference. 

Propensity matching is the opposite.

We start by looking at the users we want to analyze, which in this case is everyone who signed up for the loyalty program.

Then we look at everyone who USED to behave the same but didn't sign up for the loyalty program.

Then we measure the difference in their behaviors. 

Here's an example:
User 1 starts as an X then becomes a O.
User 2 starts as an X and stays as an X. 
We measure the difference between O and X. 

X → O (Treatment)
X → X (Control)

CMO: Woah! Basically it's a way for us to compare more apples to apples without an A/B test.

Me: Yes, exactly. 

CMO: This is awesome. Thanks for saving my a*s

Me: No problem. If you want to read more check out this article I sent to 2,800+ Marketers → https://lnkd.in/gcHsSpqy

CMO: Damn, we are not data driven at all.

CEO: What do you mean?! I've tried to be since day 1!

CMO: Because no one changes their mind based on the data.

CEO: I don't understand. We hired data scientists. We have dashboards everywhere. How are we not data-driven?

CMO: Again, because no one changes their mind based on the data.

CEO: That's not true. We look at reports every week.

CMO: Looking at data isn't the same as being driven by it.

CEO: What's the difference?

CMO: Last month, we tested two pricing models. The data showed Model B was 8% better. We still went with Model A.

CEO: We had strategic reasons for that.

CMO: Exactly. That's being opinion-driven with data decoration. Nothing wrong with it, but let's not pretend it's something else.

CEO: So what does data-driven actually look like?

CMO: It means you make the uncomfortable decision when data contradicts your intuition. Most companies can't stomach that.

CEO: Give me another example.

CMO: Our best performing acquisition channel according to dashboards? I wanted to cut it. You pushed back because "it's always worked." That's not data-driven.

CEO: But you didn't have proof it wasn't working.

CMO: Our team had a solid analysis that suggested it was plus experience. We wanted to run an incrementality test proving it out. You wanted more proof. That's the problem. Data-driven companies trust the data AND run tests to confirm hypotheses. You just wanted a narrative that felt right.

CEO: This is uncomfortable to hear.

CMO: I know. It’s not easy for me to says That's how we know we're getting somewhere. We say we want data-driven culture but actually want data-supported decisions.

CEO: What's it take to actually become data-driven?

CMO: Start saying "the data changed my mind" more than "the data confirms what I thought." Been reading Sundar's posts on this. He calls out the theater constantly.

CEO: This Sundar guy is on to something. Send me his LI profile.

CMO: Sent below 
--
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this post later

New CMO: Why do we separate brand and performance?

CEO: Because they’re different things.

CMO: How so?

CEO: Brand is long-term and hard to measure. Performance is short-term and drives revenue. We split the budget 30/70 to keep it clean.

CMO: That’s the problem. You’re acting like they work in isolation.

CEO: They don’t?

CMO: Our performance channels work better when brand awareness is high. And brand campaigns drive some immediate conversions. They’re not separate games.

CEO: So how should we think about them?

CMO: Stop asking “is this brand or performance?” Start asking “what’s the full impact of this investment?”

CEO: That sounds complicated.

CMO: It is. But pretending they’re separate is why companies think brand doesn’t work.

CEO: Explain.

CMO: We run a brand campaign. Performance metrics don’t move much immediately. You call it a failure. But we didn’t measure if it lowered our CAC or improved conversion rates on paid channels.

CEO: How would we measure that?

CMO: Geo holdouts. Test brand in some markets and not others. Then measure the impact on all channels, not just brand-attributed conversions.

CEO: Why doesn’t every company do this?

CMO: Because it’s easier to keep them in silos. Brand team reports on awareness. Performance team reports on ROAS. No one connects the dots.

CEO: We’ve been doing exactly that.

CMO: I know. Time to change it up. 
–
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit later

experiMENTAL Newsletter & Podcast  has 500 downloads! 

If you’re a consumer tech marketer looking to learn what works class teams are doing, I biasedly think this is the best place to start.

I cover everything from brand building to deep data science with leaders from Uber , Spotify , Anthropic , Wayfair , and so many more amazing companies! 

I’m now up to 10 episodes out too so you don’t need to just listen to one episode and then get sad you can’t keep bingeing. 

Start listening now: https://lnkd.in/gFGjqJ3e

Or watch : https://lnkd.in/g8VPT8MD

Here are 4 stories from my time leading Uber's Brand Data Science team that will change your mind about the impact of Brand Marketing:

All of these stories are empirical with data backed causal studies and not theoretical. 

1. DoorDash kicks our b*tt
In 2018, UberEATS and DoorDash were neck in neck in the US. 

Now? It has 66% marketshare.

The most readily obvious reasons are selection and focusing on suburbs. These are definitely primary reasons but they also outspent us on Brand Marketing early. 

They were on TV and OOH faster and bigger which is exactly where suburban audiences consume and it made them the entrenched preference.

The UberEATS team is world class but once that brand moat sets in it becomes a tough uphill battle. 

2. Needing to be always on

We ran a brand campaign in the UK where we were testing a new concept for Uber as a whole.

During the campaign, awareness went up massively. 

Think XX%. Signups did too across Rides and Eats.

But we then had to pause the campaign and within a month that awareness came plummeting back to the pre campaign base line. 

And signups? You guessed it. 📉

We then advocated for always on budget. 

3. Know where your audience is

In India, we were launching a new campaign around women's safety in Uber Auto.

The classic perf marketing channels work but not as well as other channels especially in dense cities in India and we wanted the message to be more public.

So OOH and TV were where we invested and we saw Auto trips go up by 6% within a short period. 

4. In Australia, we went all in with a campaign called "Tonight I'll be Eating" for UberEATS that featured Kim Kardashian and a well known Australian comedian. 

The campaign took off! 

We were hitting record bests and the impact was immediate cementing our place as the #1 app in Australia.

I've got a dozen more stories, but the point is as someone who was lucky enough to causally measure these campaigns, I got to see the impact of Brand at a scale that no other marketing type can touch. 

Brand campaigns have massive impact.

You just need to figure out how.
--
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this post later

Sebastian Hewing 🚀 is my next guest on experiMENTAL Newsletter & Podcast! 

He's a data and AI advisor to VC and PE-backed startups after working at some of the fastest growing startups ever with Rocket Internet SE.

On today's episode we chat about: 
• how to build data setups that empower, not babysit, your stakeholders
• a three-step framework that’s helped 40+ companies
• best traits of high-leverage your first analytics hire
• why solution-first thinking fails data teams

My favorite quotes from our conversation:

"I always implement a role that I call the super user, which
is a role that that I think is lacking in a lot of data setups and causes all kinds of problems."

"I think it's a very poetic point. The data advisor is not coming in for the data analyst they're coming for to help out another function."

"I think there's the pre AI problems and then there's the pre AI challenges and then there's the additional challenges that come with everything that's happening around AI."

Sebastian is by far one of the best data leaders I've had the pleasure of chatting with and if you're either a data leader looking to uplevel or a stakeholder looking to learn how to best implement data practices, this one's for you.

Tune in wherever you love listening to podcasts:
Listen: https://lnkd.in/ebSGnypU
Watch: https://lnkd.in/gPacUgjP
Subscribe: https://lnkd.in/gPacUgjP

CEO: Revenue is up but margins are down?!

New CMO: Yup, we’re just renting growth.

CEO: Explain. 

CMO: There’s a difference between earning growth and renting it.

When you started, CAC was $50 and customers had strong word-of-mouth. We were earning growth. Now CAC is $120 and referrals are flat. We’re renting it and paying the paid tax. 

CEO: So we’re spending more to acquire customers?

CMO: Worse. We’re spending more to acquire lower-quality customers. The easy audience is tapped out.

CEO: How do you know they’re lower quality?

CMO: Day 30 retention dropped from 60% to 45%. LTV is down 25%. We’re hitting diminishing returns on paid channels.

CEO: What happens if we keep scaling spend?

CMO: Revenue goes up. Profitability goes down. Eventually, we’re growing unprofitably, and the board pulls the plug.

CEO: So we stop spending?

CMO: No. We need to slow down marketing. We fix the unit economics first. Better onboarding, improved retention, stronger referral mechanics. We also need to invest in Brand Marketinf after. Then we scale again. 

CEO: Why can’t we do both at the same time?

CMO: Because resources, priorities, and bad growth masks the problem. If revenue is climbing, no one feels urgency to fix retention. Then one day, the channels stop working and we realize we built on sand.

CEO: How do we know when we’ve hit this wall?

CMO: When CAC is rising faster than LTV. When our payback period gets too long. That’s the signal.
–
🔔 Follow Sundar for Marketing Strategy & Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this post later

Measuring marketing is NOT about the latest and greatest tools.

A successful strategy has 4 components: 

1. Goals
2. People
3. Culture
4. Channels

Goals 
⮑ What is the business trying to accomplish? 
⮑ How does Marketing fit in?

People
⮑ What skillsets do you have?
⮑ What's the hiring roadmap look like? 

Culture
⮑ Are you willing to experiment?
⮑ Does leadership understand the risks?

Channels
⮑ What are the main channels used?
⮑ Are we launching new ones?

If you find yourself struggling to prove the ROI of Marketing, take a step back and figure out which of these 4 places is lacking. 

The most successful Marketing Science teams start at the top. 

The tools are never the answer. 

What’s the number one thing you focus on to implement an effective measurement strategy? 

--
🔔 Follow Sundar for Marketing Strategy and Science
♻️ Repost if you think others will find this useful
⬇️ Save so you can revisit this later

This visual perfectly captures what building experiMENTAL Newsletter & Podcast has felt like.

After my short stint as an FTE this year, I was super stoked to go full steam ahead on my newsletter and podcast about Marketing in Consumer Tech. 

It let me chase the joy of sharing everything I've learned in my career while having the ultimate flexibility to hang out with my kids and wife.

But then you run your family's financial model.

And reality sets in that your passion has to be a business.

Writing an article goes from shipping for fun to shipping and tracking if people opened, read, and converted. I'm microanalyzing every survey response or piece of feedback. 

It sucks the fun out of writing and I couldn't put my finger on why until I saw this chart from Peter Walker.

Simply, the largest change in happiness happens after "Initial Enthusiasm" when it becomes "Reality sets in". 

It has by far the steepest slope (1st derivative for you math nerds). 

And that's where I'm at.

It's scary. 

And the only thing that gets me through? 

Unwavering faith in my work. 

Faking it til I'm making it. 

Experimenting. Pivoting. Grinding.

And now I get why founders have this insane ability to force visions and missions because that's what's required.

Not only do I want to believe I produce the best content on Marketing Strategy & Science in Consumer Tech, but I have to. 

I have to believe my sh*t is worth reading.

I have to believe it's worth buying.

I have to believe it's worth it.

Because then everyone else will too.

Love, 
Sundar

Dear post reader, 

Happy Thanksgiving from 0-3 weeks ago! 

I can not control when this surfaces, but I hope you have fully recovered from the binge eating, career sh*t talking, and your football team’s loss. 

I also hope that you were able to celebrate with everyone you were thankful for and that you recognize the good things going on in your life even if it’s surrounded by some other bad. 

Happy Thanksgiving! 🍽 🦃 🥘 🌽 

Sincerely,
Sundar

My next podcast guest is Mattia Santin, Senior VP of Marketing at Contentsquare (ex Hotjar, Uber)! 

On today's episode we chat about: 
• the career leaps that shaped his approach 
• what happens when CMOs get their hands dirty 
• why too many marketing teams mistake motion for progress 

and much more.

My favorite quotes from our conversation:

"The first thing I do is spend as much time as possible talking to people"

"You have to go down, into the trenches also in the numbers and understanding, making your own assessment of what, what you're seeing"

"Everybody was pushing me to spend more. It, it didn't feel right, in part because I never had that experience, in part because I didn't feel comfortable in our measurement and our ability to understand truly ROI."

Mattia is a brilliant strategist AND executor knowing how to drop in and build impactful Marketing teams. If you’re looking to get a leg up on how to navigate strategy, scale, or structure inside a fast-growing company, this one’s for you.

Tune in wherever you love listening to podcasts:
Listen: https://lnkd.in/gFGjqJ3e
Watch: https://lnkd.in/gueY5bYq
Subscribe: https://lnkd.in/gxR5c4Xd

‼️Launching accelerator: babY Combinator

Focused on seed and preseed stage for founders aged 7 and below. 

The new generation is born with AI from day 1 and understand the technology better than anyone.

In addition, when they build products for their peers, that demographic will have the highest LTV with an average of 75-80 years ahead of them. 

The math is undeniable. 

There of course will be some challenges:
- Additional facilities for diapers
- Cofounder tantrums 
- Naps 

But we believe we can over come them and create the next generation of truly generational founders. 

It is never too young to ride the AI wave. 

DM me if you want to be an LP. 

If you’re sold, skip the convo and Venmo me

Product Growth Analyst

Poshmark

Lead Product Analyst

➡️ Collaborate closely with Product Managers at Poshmark to improve Seller Experience & Live Show Launch & Features:
✔️ Ran 30+ Experiments on the Seller Experience side (A/B and Pre-post analysis)
✔️ 25% successful experiments that led to an increase of 3-4% of GMV
✔️ Led the implementation of mweb listings on the platform, increasing the number of listers by 18k/month, 0.3% of total listers.
✔️ Led Live Shows Launch and GTM effort working cross functionality with data eng and executives

Senior Product Analyst

Tophatter

➡️  Perform detailed analyses to distill key insights & drivers for retaining the customer base
✔️ Designed & ran more than 50+ experiments to measure the impact of new features to improve buyer and seller retention
✔️ Led, designed and built a set of solutions to address buyer retention, including usage of credits and personalized push notification that improve attendees retention by 25% and payer retention by 15%

➡️  Improve platform health with automation in User Acquisition and Supply 
✔️ ROAS optimization for multiple channels spend allocation resulting in a 50% increase in D14 and D28 ROAS
✔️ Improve Seller Funnel from Acquisition to Seller Retention that resulted in a 100% increase in retention and 35% in LTV (Spend per seller)

Growth Analyst

I love Data. And I love writing. I love writing about Data.
Open for brand collab.

(still figuring out the rest)

Amney Mounir

You’re not the problem — the job market is.

I keep seeing “junior” roles asking for:
- 3+ years of SQL & Python 
- A/B testing knowledge
- Experience managing stakeholders

That’s not junior.
That’s companies filtering candidates.

They wanna pay less for more experience.
And because the market is benefiting them.

 They set the rules.
 And juniors - the least experienced - get stuck.

But here’s my recommendation:

If you’ve got strong SQL, good business sense, and can communicate insights clearly—you are enough.

Don’t let job descriptions fool you into thinking otherwise.

And if you cannot land that job, get unpaid volunteering experience, do projects for free, get that experience and then start applying.

Here are few resources to help you build these skills:
📕 The Data Book: https://lnkd.in/ei8Sp6Qw
📘 Data Roadmap: https://lnkd.in/eYpSaKMx

———

👋 I'm Amney, follow me for more Daily Tips
♻️ Repost this if you found it useful!

Grow | LinkedIn

li_at

JSESSIONID

Data Analytics & AI @ Meta

SQL is 𝘴𝘵𝘪𝘭𝘭 the king of data.

People keep saying “SQL is dead" but the reality is:

- Every interview still has a SQL round
- SQL is still the main tool to pull data
- SQL is still the most reliable language
- SQL isn’t a “nice-to-have", it's a must

Here are few resources for you to get started:
📕 Interview Book: https://lnkd.in/ei8Sp6Qw
📘 Roadmap: https://lnkd.in/eSCDkmAU

———
♻️ Repost this if you find it insightful!
👋 Follow Amney for more Daily Tips

When I was trying to break into data,

I didn’t have mentors. I didn’t know what to study.
And honestly… I spent months feeling completely lost.

Doing projects helped me a lot to build my skillsets.

Here are in my opinion the best Github repos in 2025:

1) awesome-data-analysis

𝗟𝗶𝗻𝗸: https://lnkd.in/eVMPj3Nh
𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: A comprehensive list of SQL, Python, ML, and analytics resources — a perfect starting point for anyone building a full data foundation.

2) data-scientist-hanbook

𝗟𝗶𝗻𝗸: https://lnkd.in/eMzixkG3
𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: A curated, minimalist handbook from Andres that tells you exactly what to learn (and what to skip) for modern data roles.

3) cookiecutter-data-science

𝗟𝗶𝗻𝗸: https://lnkd.in/eBdvtNkE
𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: An industry-standard project template to structure your analytics and ML projects the same way professionals do.

4) data-science-template

𝗟𝗶𝗻𝗸: https://lnkd.in/erbER7XP
𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: A clean, reproducible project setup that teaches you how to build maintainable, production-ready data projects.

5) ML-From-Scratch

𝗟𝗶𝗻𝗸: https://lnkd.in/excyjzXn
𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: Clear, from-scratch implementations of key ML algorithms — perfect for understanding models and explaining them in interviews.

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

What it feels to be a Data Analyst in 2025:

Saying “I’m learning ML” will NOT get you hired.

Showing 𝘪𝘮𝘱𝘢𝘤𝘵𝘧𝘶𝘭 projects you’ve built will.

Especially with code, models, and results.

Here are 5 projects to make your portfolio stand out as a Data Scientist or ML Engineer: 👇

1. 𝐏𝐫𝐞𝐝𝐢𝐜𝐭 𝐂𝐡𝐮𝐫𝐧 𝐰𝐢𝐭𝐡 𝐗𝐆𝐁𝐨𝐨𝐬𝐭 

Use customer data to predict churn.
Clean data → feature engineering → tune your model.

Link: https://lnkd.in/eDcHXdy2

2. 𝐑𝐞𝐚𝐥 𝐀/𝐁 𝐓𝐞𝐬𝐭 𝐄𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐨𝐧

Run a simulated or real test.
Use t-tests or Bayesian stats to back up your insights.
Explain trade-offs clearly.

Link: https://lnkd.in/eZ2ZDr7z

3. 𝐓𝐢𝐦𝐞 𝐒𝐞𝐫𝐢𝐞𝐬 𝐅𝐨𝐫𝐞𝐜𝐚𝐬𝐭𝐢𝐧𝐠 

Use Prophet or ARIMA to forecast revenue.
Clean the trend + seasonality. Plot actual vs. predicted.

Link: https://lnkd.in/ePFvMKkx

4. 𝐍𝐋𝐏 𝐒𝐞𝐧𝐭𝐢𝐦𝐞𝐧𝐭 𝐀𝐧𝐚𝐥𝐲𝐬𝐢𝐬 

Scrape reviews. Clean and vectorize text.
Train a classifier (LogReg or BERT).

Link: https://lnkd.in/e4ncGEf5

5. 𝐑𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐞𝐫 𝐒𝐲𝐬𝐭𝐞𝐦 (𝐀𝐢𝐫𝐛𝐧𝐛 𝐒𝐭𝐲𝐥𝐞)

Use collaborative filtering or content-based approach.
Matrix factorization, similarity scores, etc.

Link: https://lnkd.in/etiyDarf

———

♻️ Repost this if you found it insightful!
👋 Follow Amney for more Daily Tips

I analyzed 200+ Machine Learning job posts. 

And here is what I found:

Over 90% of ML interviews are 𝘤𝘰𝘯𝘤𝘦𝘱𝘵𝘶𝘢𝘭.
And don’t need you to code. 

The interview will assess how you think, explain, and reason through trade-offs.

Here are few questions you can expect:

1) How do you decide between Logistic and Linear Regression?

2) What’s the role of regularization in model performance?

3) How does Gradient Boosting differ from Random Forest?

4) Why does feature scaling matter for SVMs?

5) How do you evaluate models on imbalanced data?

If you have no clue how to answer these.

𝘐𝘵’𝘴 𝘧𝘪𝘯𝘦, 𝘥𝘰𝘯’𝘵 𝘸𝘰𝘳𝘳𝘺!

You can get better with 𝟑𝟔𝟓 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐜𝐞

They are offering free 15-day access (Nov 6–21) to 115+ AI & Data Science courses and real-world projects.

📘 Access for FREE: https://bit.ly/4nHhCyA

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips
#AIandDataLearningChallenge

People will only show up when you get the job. 

Not when you’re stuck job hunting.

This process is hard, and you feel lonely, you're trying to stay motivated, trying not to take things personally, trying not to give up. 

Here’s what I wish someone told me:

- Rejection doesn’t mean you’re not good enough. 
- Your "yes" is coming soon
- Stay consistent

And I'm sure, 𝗬𝗢𝗨 will get there!

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

Most aspiring Data Analysts stop at learning hard skills.

And that’s exactly why they get stuck.

They'll learn SQL, Python, Excel, Tableau...

But this is the part nobody talks about:

Being a great Data Analyst requires TWO skillsets.
And most people only build one.

Great Analysts can:

1) Solve open-ended, messy problems
2) Drive decisions from ideation → execution
3) Communicate insights people understand
4) Collaborate and build strong relationships

Don't get me wrong, I want you to master hard skills.
But invest twice as much in your soft skills!

That’s what will make you stand out in 2025.

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

What differentiates Junior and Staff Data Analysts?

It’s mostly 3 things:

- Scope
- Influence
- Ownership

➡️ Junior Data Analyst:

- Works on small, well-defined problems
- Primarily executes tasks given to them
- Doesn’t influence decisions much

➡️ Senior Data Analyst:

- Owns a medium scope of work
- Have an opinion and influence discussions
- Shape direction rather than just following orders

➡️ Staff Data Analyst:

- Operates at a high scope, often across multiple teams
- Works as a consultant, delegates, and drives decisions
- Recognizes when things are off track and proactively corrects course

Where do you stand? 👇

———
♻️ Repost this if you find it insightful!
👋 Follow Amney for more Daily Tips

Acing your data interview takes more than technical skills.

You need to be able to:

1) Communicate clearly your projects
2) Being able to solve open-ended questions
3) Come up with the right metrics to measure
4) Be great at SQL (the most asked technical skill)

Being able to do these 4 things well, will put you in the best spot to land your next Data Role.

In the next year, I will expect AI being part of the interview process for Data Analysts.

Here are few resources to get you started:
📗 The Interview Book: https://lnkd.in/ei8Sp6Qw
📕 Interview Questions: https://lnkd.in/eHEb4u3P

———

♻️ Repost this if you find it insightful!
👋 Follow Amney for more Daily Tips

I’d have NEVER landed a job at Meta with 𝘫𝘶𝘴𝘵 SQL.

Here’s exactly what I needed to get there: 

1) Success Metrics

- Understand product goals
- Define your north Star
- Break it into input metrics
- Slide and dice them further

2) Investigation

- Clarify the drop
- Dig into internal factors
- Mention external factors
- Summarize all your findings

3) Product Ideation

- Clarify the goal
- Lay down the User Funnel
- Generate Ideas for each funnel step

4) Behavioral

- Use the S.T.A.R method
- Focus on metrics for your results & impact!

With all of that, you'll be safe out there.

You got this!

———

♻️ If that was useful, Repost bc people need to know!
👋 Follow Amney for more Daily Tips

Here is how I would prioritize my time to prepare for my next Data Analyst interview:

Junior (0-2 years)
- 70% SQL
- 20% Case Study
- 10% Behavioral

Mid-Level (2-4 years)
- 50% SQL
- 30% Case Study
- 20% Behavioral

Senior (5-8 years)
- 50% Case Study
- 30% Behavioral
- 20% SQL

Staff (8+ years)
- 50% Behavioral
- 40% Case Study
- 10% SQL

If you still don't know how to build your business sense, I highly recommend this free event from Leland, this will def set you apart if you start building this skill!

📕 Register here:  https://lnkd.in/eCgvyvnG

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

This is all you need to know in Excel as a Data Analyst:

This is from my experience the things I've used in Google Sheets.

I never used Macros or other fancy stuffs.

If you are able to know these 7 things you're good to go:
- Pivot Tables
- Math Functions
- VLOOKUP
- Index Matching
- Charts
- Logical Functions
- HLOOKUP

Here are my favorite resource to get you ready:
📕 Data Book: https://lnkd.in/ei8Sp6Qw

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

A Master’s Degree or PhD will 𝐍𝐎𝐓 save you anymore.

This is because the job market is changing faster than universities can adapt.

You don’t need more degrees.
You need 𝘯𝘦𝘸 skills built for this era.

Here’s what I’d focus on:

✅ Learn to compete with AI, not against it
✅ Sharpen your storytelling and business sense
✅ Start building — use AI tools like Replit or Lovable

I don’t know what the future of data jobs will look like.

But I know one thing — it won’t be the same.

So please, don’t hide behind degrees or certificates.
Keep adapting. Keep experimenting.

That’s how you stay relevant.

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

A Master’s degree isn’t a golden ticket anymore.

“Do I really need one to get a data job?”
“Will it help me break into tech?”
“Is it worth all these debts?”

Before, having a Master’s 𝘸𝘢𝘴 an advantage.
But, today it's 𝗼𝗽𝘁𝗶𝗼𝗻𝗮𝗹.

You don’t need more degrees.

You need more skills:
- SQL
- Story-telling
- Product Sense
- Real Experience

That’s what gets you hired.
𝘕𝘖𝘛 a piece of paper that cost $60,000.

If you’re going to do it, do it for the right reasons:
→ immigration
→ networking
→ career break

But don’t do it because you think it 𝘨𝘶𝘢𝘳𝘢𝘯𝘵𝘦𝘦𝘴 a job.

Those days are gone.

———

♻️ Repost this if you found it useful!
👋 Follow Amney for more Daily Tips

taplio

The Battle

Sundar Swaminathan

Amney Mounir

Followers

Posts

Engagements

Likes

Performance ScoreSundar Swaminathan's LinkedIn Performance

Performance ScoreAmney Mounir's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Sundar Swaminathan

Amney Mounir

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Sundar Swaminathan's LinkedIn Performance

Performance ScoreAmney Mounir's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreSundar Swaminathan's LinkedIn Performance