Podcast on PepsiCo: Advanced Analytics for Growth

Kapitoly

Úvod

Konec průměrů

Od tabulek k umělé inteligenci

Záhada postmixového nápoje

Jak udělat statistiku zábavnou

The Demand Accelerator

Introducing PEPWORX

From Transactional to Collaborative

The Proof is in the Ranking

The Four Pillars of Power

When the Game Changes

The Demand Accelerator

Cheetos from the Grave

An Analytics Arms Race

The One-Page Vision

Not a Think Tank

Three Big Goals

The Oatmeal Example

Hunger for Problems

Warehouse vs. Lake

Building a Next-Gen Platform

The Power of Shopper Data

Who Pays for Research?

Patrons vs. Sponsors

Přepis

Noah: Tady je otázka, která na zkoušce z byznysu nebo marketingu zaskočí 80 % studentů: Jaký je skutečný rozdíl mezi sledováním prodejů a skutečnou maloobchodní analytikou? Zní to podobně, že? Ale pochopení tohoto rozdílu je hranicí mezi průměrnou známkou a jedničkou. A my vám přesně ukážeme, jak na to.

Sara: Přesně tak. Není to jen o tom, co se prodalo. Jde o to pochopit, *proč* se to prodalo, kdo to koupil a co si koupí příště. A technologie, která za tím stojí, je fascinující.

Noah: Zůstaňte s námi a už nikdy se na nákupní uličku nebudete dívat stejně. Jste u Studyfi Podcast.

Noah: Dobře, Saro, začněme od základů. Co přesně je „maloobchodní analytika“ a proč jsi zmínila, že to není jen sledování prodejů?

Sara: Skvělá otázka. Dříve se firmy jako PepsiCo dívaly na data z velké dálky. Řekněme, že se podívali na prodeje v celém městě nebo PSČ a řekli si: „Dobře, v této oblasti se prodává hodně Coly Zero.“

Noah: To zní celkem logicky. Co je na tom špatného?

Sara: Problém je v tom, že je to průměr. A průměry skrývají detaily. Co když je v jedné ulici obchod, kam chodí hlavně studenti, a přes ulici jiný, kam chodí rodiny s dětmi? Jejich nákupní zvyklosti budou naprosto odlišné.

Noah: Takže i když jsou od sebe jen pár metrů, potřebují jiné produkty na regálech?

Sara: Přesně tak! A to je ten skok k moderní analytice. Jde o granularitu. PepsiCo si uvědomilo, že musí rozumět lidem. Co je motivuje? Jaké jsou jejich zvyky? Začali se dívat na data na úrovni jednotlivých obchodů, dokonce i v okruhu pár bloků, aby našli skryté příležitosti.

Noah: Takže už žádné univerzální řešení pro všechny. To je docela velká změna myšlení.

Sara: Obrovská. Umožnilo jim to dělat stejné věci – jako je uspořádání zboží – ale mnohem chytřeji. A pak je to povzbudilo dělat věci úplně jinak.

Noah: Dobře, ale mluvíme o obrovském množství dat. To se nedá jen tak hodit do Excelu, že?

Sara: To rozhodně ne. To je další klíčový bod. Zpočátku analytici v PepsiCo používali hlavně tabulky a vytvářeli hypotézy. Ale jak dat přibývalo, potřebovali silnější nástroje.

Noah: Tady přichází na řadu ta „pokročilá analytika“?

Sara: Ano. Přešli k statistickému modelování, které jim umožnilo předpovídat budoucí chování. A nakonec se dostali ke strojovému učení.

Noah: Strojové učení. To zní složitě. Jak to funguje v praxi?

Sara: Představ si to jako superchytrého asistenta, který dokáže v datech najít vzorce, které by člověk nikdy neviděl. Dokáže analyzovat tisíce proměnných najednou – od počasí přes místní události až po demografii okolí – a vytvořit neuvěřitelně přesné předpovědní modely.

Noah: Takže stroj v podstatě říká: „Hej, v tomto konkrétním obchodě, v úterý, když svítí slunce, si lidé s největší pravděpodobností koupí dietní limonádu“?

Sara: Přesně! A to s mnohem větší přesností, než bychom dokázali odhadnout. Je to o přechodu od dohadů k rozhodnutím založeným na datech.

Noah: To je teorie, ale máš nějaký konkrétní příklad, kde to opravdu změnilo hru?

Sara: Mám perfektní. Představ si čerpací stanici, kde si lidi kupují kelímek a sami si čepují nápoj z postmixu. Prodejce ví, že prodal kelímek, ale absolutně netuší, co si do něj zákazník načepoval.

Noah: Aha, takže neví, jestli má objednat víc Pepsi nebo třeba 7UP sirupu. To je problém.

Sara: Přesně ten problém. Takže datový tým PepsiCo udělal něco chytrého. Vzali data o tom, kolik galonů sirupu jednotlivé obchody spotřebovaly – a to nejen od PepsiCo, ale i od konkurence.

Noah: Dobře, takže teď vědí, co se pije. Ale jak zjistí, *proč*?

Sara: A tady přichází ke slovu strojové učení. Vytvořili soubor modelů, které analyzovaly vlastnosti daného obchodu a jeho zákazníků a propojily je se spotřebou sirupu. Zjistili tak, jaký typ obchodu prodává jaký typ nápoje.

Noah: Takže mohli přijít za majitelem a říct: „Váš obchod se podobá profilu X, a pro tento profil je nejdůležitější mít v nabídce tyhle tři příchutě.“

Sara: Přesně tak! Dali jim seřazený seznam nejdůležitějších sirupů pro různé typy obchodů. Ten prodejce to pak zavedl ve více než tisícovce obchodů a začal měřit dopad. Už to nebylo hádání.

Noah: Tohle je super, ale jak přesvědčíš manažery prodeje, kteří jsou zvyklí spoléhat na svou intuici, aby věřili nějakému algoritmu?

Sara: To je skvělá pointa. Většina lidí nesnáší statistiku. Ale zjistili, že ji začnou milovat, když jim usnadní práci. Klíčové bylo jim vše vysvětlit.

Noah: Takže žádná černá skříňka, kde jen vypadne výsledek?

Sara: Vůbec ne. Provedli je celým procesem. Začali jim představovat pojmy jako z-skóre. Když se ptali, proč je to lepší, vysvětlili jim, že teď nevidí jen průměrné prodeje, ale i vrcholy a pády. Vidí celý obrázek.

Noah: Takže v podstatě udělali ze statistiky superhrdinu, který odhaluje skryté pravdy v jejich datech.

Sara: Přesně tak! Když lidé pochopili, *proč* a *jak* se k výsledku dospělo, začali těmto novým nástrojům důvěřovat a aktivně je využívat.

Noah: Klíčovým poznatkem tedy je, že data sama o sobě nestačí. Musíte je umět srozumitelně interpretovat a ukázat jejich reálný přínos. To je skvělý přechod k našemu dalšímu tématu...

Noah: So, that makes sense. But how does a giant company like PepsiCo even begin to organize that much information?

Sara: That’s the multi-billion dollar question, Noah. And for PepsiCo, the answer started back in 2015.

Noah: 2015? What happened then?

Sara: They created a centralized team called the PepsiCo Demand Accelerator. Think of it as an internal startup... a super-smart group focused on one thing: using data to innovate.

Noah: So, not just counting bags of chips, then.

Sara: Not even close. Over four years, this team built an advanced analytics platform. They gathered a massive, anonymized dataset on shopper behavior.

Noah: Anonymized is the key word there, I hope. So what did they do with all this data?

Sara: They created information solutions to help their own sales teams, but here's the game-changer... they also created tools for their retail customers.

Noah: Their customers? You mean like the grocery stores?

Sara: Exactly. By 2019, they formally packaged these tools into a suite of data analytics capabilities called PEPWORX.

Noah: PEPWORX. Sounds like a high-tech energy drink.

Sara: It's just as powerful! PEPWORX helps retailers figure out what to put on their shelves and where. It helps them optimize store space and manage their marketing dollars more effectively.

Noah: Okay, let me give you an example. Does PEPWORX tell a store in Texas to stock more spicy chips than a store in Minnesota?

Sara: That's the core idea, but on a much more granular level! It helps them unlock growth by understanding shoppers at a local, even a single-store, level.

Noah: You said this was a game-changer. Why does helping the retailer matter so much to PepsiCo?

Sara: Because it completely transformed their relationships. It shifted them from being transactional to being collaborative.

Noah: What's the difference?

Sara: A PepsiCo executive used a great clock metaphor. A transactional relationship is between 1:00 and 3:00 on the clock. It's not deep, it's mostly about price.

Noah: Just selling stuff. Got it.

Sara: Right. But as you move around the clock, the relationship becomes collaborative. You’re co-creating solutions. With these data tools, PepsiCo moved way past 3:00 with many of their partners.

Noah: So they're not just a supplier anymore. They're a consultant... a partner.

Sara: Precisely. Their goal became what they call a “three-audience win.” The shopper wins, the retailer wins, and PepsiCo wins. Everyone benefits.

Noah: That sounds great in theory, but did it actually work? Did retailers notice?

Sara: They absolutely did! There’s a very visible industry survey called Kantar's PoweRanking. For four straight years, retailers voted PepsiCo their number one partner in North America.

Noah: Four years in a row! That’s incredible. That's the payoff right there.

Sara: It is. It proves that sharing insights and acting like a true partner is a winning strategy. They even formalized this with something called Joint Business Planning, or JBP, where they sit down with retailers annually to plan for mutual growth.

Noah: So the key takeaway here is that using data to help your partners succeed is actually the smartest way to help yourself succeed.

Sara: That's it exactly. It’s a powerful lesson in modern business. This wasn't just about tech; it was about building trust through data.

Noah: It really redefines what a partnership can be. Now, this is a massive scale, but it makes you wonder how smaller businesses can apply these same principles...

Noah: So that idea of a legacy company needing to reinvent itself is huge. And it brings us to a perfect case study... PepsiCo.

Sara: Exactly. We all know them. Lay's, Doritos, Gatorade. By 2019, they were a 65 billion dollar company. They were the definition of success.

Noah: But that's the problem, right? When you're that big, how do you keep getting bigger?

Sara: That was the fundamental challenge. Michael Lindsey, an exec at Frito-Lay, put it perfectly. How do you grow at the pace you're used to, without just... spending more money than you're making?

Noah: So what was their old playbook? How did they get so massive in the first place?

Sara: It rested on four key pillars. First, they had iconic brands. We're talking 22 different brands that each made over a billion dollars a year.

Noah: Wow. Okay, that's a strong start. What's next?

Sara: Second, their reach was insane. They had the largest fleet of trucks in the United States. Bigger than UPS! They could get their chips and drinks into hundreds of thousands of stores, everywhere.

Noah: That's a huge advantage. What were the other two?

Sara: Third was a strong value proposition. They were masters of manufacturing at scale, which kept their costs low and prices competitive. And fourth, of course, was a massive market presence through huge ad campaigns.

Noah: So... what broke? Why did that incredible formula start to fail in the 2010s?

Sara: The world just changed around them. Suddenly, private labels from grocery stores got really popular. And shoppers started looking for things that felt new and fresh.

Noah: Ah, the rise of the kale chip!

Sara: Exactly! People wanted snacks made from chickpeas, seaweed, things PepsiCo wasn't known for. Plus, new online stores were popping up, and their key demographic, the baby boomers, were aging.

Noah: So their whole playbook was designed for a game that wasn't being played anymore. What did they do?

Sara: They decided to get smart. Really smart. They realized their size and product variety could still be a competitive advantage, but they had to use it differently.

Noah: How so? What was the new plan?

Sara: They decided to use data and advanced analytics to find what they called 'latent consumer demand'. Little pockets of opportunity hidden everywhere.

Noah: Okay, 'latent consumer demand' sounds like a textbook term. What does that actually mean?

Sara: Fair enough. Think of it this way: instead of trying to sell Doritos to everyone, they wanted to figure out which specific corner store in which specific neighborhood should stock the Flamin' Hot Limon flavor. It's about getting super granular.

Noah: And how did they do that? Did they just create a new department?

Sara: Pretty much! They created the 'Demand Accelerator,' or DX. It combined teams that looked at shopper analytics, in-store space optimization, and marketing, all powered by data.

Noah: Okay, I need a concrete example. Show me this in action.

Sara: The perfect one is Cheetos Asteroids. Remember those?

Noah: Vaguely! They were discontinued a long time ago, right?

Sara: Yep, back in the early 2000s. But people were still talking about them online, even starting petitions to bring them back. The thing is, it was a niche audience. A big, general relaunch would probably flop.

Noah: So what did the DX team do?

Sara: This is the cool part. They used social media data to pinpoint the exact geographic locations where these conversations were happening. Then, they overlaid that with shopper data to find the specific convenience stores those fans were most likely to visit.

Noah: No way. So they didn't just bring it back... they brought it back only to the stores where they knew the die-hard fans were.

Sara: Precisely! They invested in targeted messaging just for those areas and stocked the shelves in just the right places. And the result? They couldn't keep them in stock! It was a massive success because it was so precise.

Noah: That's brilliant. It's not about shouting to the whole country; it's about whispering the right message in the right person's ear. What a powerful shift.

Sara: It absolutely is. It shows how even a giant can learn to be nimble. Now, that idea of using data for precision targeting isn't just for snacks. It's also transforming the world of finance, which is where we're headed next.

Noah: So that’s how these big data concepts work in theory. But it’s way more interesting to see how a real giant, like PepsiCo, put it all on the line.

Sara: Exactly. And for them, it was a race. In the mid-2010s, they saw big retailers investing billions in analytics to understand shoppers. PepsiCo knew they had to keep up.

Noah: They were worried about getting left behind on the shelf, literally.

Sara: Pretty much! A senior VP, Jeff Swearingen, said they had “pockets of strength” in data, but they weren't connected. It was like having great musicians in different rooms, but no orchestra.

Noah: So how did they convince the top bosses to build this orchestra?

Sara: With a really powerful, simple idea. Swearingen’s team presented a single page to the Executive Committee. It just described a future where they could use data to give their retail partners the perfect product mix for their customers.

Noah: Just one page? That’s bold. I like it.

Sara: It worked! The committee was sold. So in 2015, they launched the PepsiCo Demand Accelerator, or the DX. They pulled about 180 people from all across the company to form this new super-group.

Noah: A super-group... I'm picturing data scientists in capes.

Sara: Not quite, because they had a very clear mission. Swearingen told them, “We’re not a think tank... we’re a commercial team.” They were expected to be self-funding and deliver real sales growth.

Noah: Wow, so the pressure was on from day one. This wasn't just an experiment.

Sara: Not at all. He even called the first members “founders,” not just employees. He wanted them to feel like pioneers building something crucial for PepsiCo's future. It was a massive cultural shift.

Noah: So they built a dedicated, high-stakes team to connect all their data knowledge. That’s the edge they needed. So what happened next? How did the DX actually start changing the game for them on the ground?

Noah: And that's a perfect lead-in to PepsiCo's Demand Accelerator, or DX. They weren't just collecting data—they had a mission.

Sara: They really did. The leader, Jeff Swearingen, gave his team three huge goals. First, change internal strategies. Second, boost their reputation with retailers. And third, get those retailers to use PepsiCo’s new insights to make better decisions.

Noah: That sounds like a lot of freedom for a team in such a massive company. Usually, the goals are much narrower.

Sara: Exactly. They were told the main outcome was to drive revenue, but how they got there was wide open. So they started with what they called “quick wins.”

Noah: A quick win sounds great. How does that work in practice?

Sara: Let me give you an example—Quaker Overnight Oats. The DX team used their new analytics to identify 24 million U.S. households that were likely to buy it, based on things like jobs and healthy eating habits.

Noah: So they basically found everyone who's too busy to make a proper breakfast? I feel seen.

Sara: Pretty much! Then they helped target marketing to stores near those households. And here's the surprising part... that approach was credited for eighty percent of the product's sales growth in the first 12 weeks.

Noah: Eighty percent? That’s not a quick win; that’s a knockout. So they just kept doing that?

Sara: They did that, and they also started hunting for bigger, transformational projects. One of their leaders said they were just constantly hungry, pitching ideas to anyone internally with a challenging problem.

Noah: I love that attitude. Not waiting for problems to come to them.

Sara: Right. But they were strategic. They only invested time if the solution would unlock sales growth *and* inspire their retail partners to actually take action. If a retailer wouldn't act on the insight, it wasn't worth it.

Noah: That makes so much sense. It's not just about being smart; it's about creating real-world impact. So, how was this 'hungry' team actually structured to pull all of this off?

Noah: So that's the “what” and the “why.” But the big question is… how does a massive company like PepsiCo even begin to manage all that information?

Sara: That's the million-dollar question, Noah. It's not like they just dumped it all into one giant folder on a computer.

Noah: I'm guessing not. So what was their first step?

Sara: They started back in 2012 by building what's called an Enterprise Data Warehouse. Think of it as a super organized, highly structured library where all the data is curated and trusted. It was great for clear, specific questions.

Noah: But I'm sensing a “but” coming...

Sara: You got it. That warehouse was too slow for the really advanced stuff, like predictive analytics. It wasn't built for exploring or asking brand new questions. So, they considered building a “data lake” to complement it.

Noah: A data lake? Sounds refreshing.

Sara: It can be! But they were careful. They didn't want it to become a “data swamp” — just a messy pile of unusable information. So they waited for the right project to come along.

Noah: And that project was the Digital Experience, or DX, team we mentioned?

Sara: Exactly. The DX team's needs became the perfect use case. Together with IT, they built what they called the “Next-Generation Advanced Analytics Platform.” It was cloud-based, which was a huge deal.

Noah: Why was being on the cloud so important here?

Sara: It meant they could pull in massive, diverse datasets from both inside and outside the company. And, it allowed them to securely share insights with their retail partners. They rolled the whole thing out in just six months.

Noah: Okay, so they built this powerful new platform. What kind of data was flowing into it?

Sara: This is where it gets really impressive. The platform gave them access to what they called “Most Valuable Shopper” data. We're talking a 20-terabyte database with over 100 million shopper records.

Noah: Wow. And what did they know about these shoppers?

Sara: Almost everything! They had over 1,800 different attributes for each record. They pulled data from everywhere — census data, weather patterns, and anonymized shopper profiles to build this incredibly rich picture of who was buying what, and why.

Noah: So they're not just looking at what I buy, but probably the weather forecast for when I'm likely to buy it. That's wild. Now, once they have all this data organized, how do they actually start making predictions with it?

Noah: Alright, so that covers how research gets *done*. But let's talk about a huge piece of the puzzle... who actually pays for it all?

Sara: An excellent question, Noah. It's rarely just the university's own money. We're talking about research sponsorship. This is how major research centers, like MIT's Center for Information Systems Research, or CISR, can do their amazing work.

Noah: Sponsorship... so like when a brand sponsors a YouTuber or a sports team?

Sara: Exactly! Think of it that way. Companies fund the research because they want access to the insights. It's a win-win. The center gets funding, and the companies get cutting-edge knowledge to stay ahead.

Noah: So it’s not just one big pile of money. How do they structure it?

Sara: Great question. They often have different tiers. For MIT CISR, there are two main levels: Research Patrons and Research Sponsors. It’s all about the level of access and involvement.

Noah: Okay, break that down for us. Patrons versus Sponsors. What's the difference?

Sara: Think of it like a video game subscription. Sponsors get the premium pass. They get a lot of great content and access. But Patrons? They're on the next level up. They get the all-access, backstage pass. They're more deeply involved.

Noah: So Patrons are like the VIPs who get to hang out with the band after the show?

Sara: That's a perfect way to put it! They have a closer relationship and contribute more, so they get more in return. Companies like Microsoft and Cognizant have been Patrons, while a wider list like Bayer and Chevron have been Sponsors.

Noah: That makes so much sense. And knowing this now gives you a huge leg up when you see these terms in the real world. So, that's our final topic!

Sara: It is! To recap, we've walked through choosing a research topic, finding credible sources, and now, understanding who funds it all. The key takeaway is that great research is a collaborative effort.

Noah: Absolutely. We hope this gives you the confidence to tackle any research project. You've got this! Thanks so much for joining us, Sara.

Sara: It was my pleasure, Noah. Happy researching, everyone!

Noah: And that's a wrap on this episode of the Studyfi Podcast. We'll see you next time.