Life is full of wicked problems. A problem is wicked when it’s not solvable in the mathematical sense of the word: requirements are contradicting, solutions aren’t repeatable even in slightly different contexts, and the problem changes as you’re solving it.

It sounds terrifying, right? Reality is that the vast majority of life decisions are addressing wicked problems: selecting your school, your partner (in business or life), deciding where to live, what kind of career to pursue.

Similarly at work, we frequently deal with wicked problems: from which features to build and what kind of marketing campaign to push, to what architectural pattern to apply to a software problem. A common theme with wicked problems is that the effects of these decisions typically lag months or years making it very hard to get a clear understanding of cause and effect.

Reality we live in is complex and volatile (and probably getting more volatile due to globalization and technology, but more on that some other time). We all need predictable and reproducible models to deal with this world. Hence all the books and frameworks on how to raise kids and how to build great products and companies.

All these books and frameworks suffer from survivor bias. This reminds me of the story about a national coin-flipping contest Warren Buffet wrote about 40 years ago: if 300 million people start flipping coins, after 20 rounds there will be 200 people who got heads 20 times in a row. At least some of them will start writing books about their coin flipping techniques.

There are many examples to pick, but I’ll focus on two that are fairly prominent and popular in the tech circles:

• Eric Ries (of Lean startup fame) was much more successful as a book author and Silicon Valley influencer than as a founder. Lean startup likely has far more readers than IMVU has MAUs 🙂
• Marty Cagan (of Silicon Valley Product Group fame) is a successful tech executive with a long career in the industry, but judging by his Linkedin hasn’t started a globally successful product company.

Most of these books (and the two quoted above especially) are full of useful ideas and frameworks. I use them on a weekly basis. But reality is messy. Data from experiments is rarely so clear-cut to make product decisions easy. A/B tests only measure first order effects. The predominant factor appears to be luck after all — being at the right place in the right time, with the right proposition.

In that world, the main thing we can do is to make it easier for luck to find us — increase optionality by making asymmetric bets as often as we can. Eventually you do get lucky. And I fully appreciate the irony of a guy working for a gambling company preaching you luck.

Peter Drucker, supposedly (*)

I always found this frequently cited quote a bit misleading. It implies a dilemma that’s in my opinion fake: no strategy will be executed effectively without a great culture evolving it and the main measure we have of corporate culture success is strategy execution under competitive conditions. The issue is nuanced and that nuance matters. My view is that in the long run, culture actually *is* strategy. More on that below.

Why do firms exist in the first place?

Ronald Coase in his 1937 article The nature of the firm (the one that got him the Nobel prize) views the firms primarily as a cost arbitrage phenomenon: when costs of doing a task internally are lower than contracting the same task on the external market, entrepreneurs hire people. This implies that firms will grow until the arbitrage opportunity no longer exists: when firms become so inefficient that it’s cheaper to contract out the firm’s tasks, firm’s growth stops.

But the firm has to be more than a mere aggregation of all the tasks being done inside it. Emergence of the internet and remote work should have decreased the cost of outsourcing firm’s tasks which should in turn lead to modern companies getting smaller than historical juggernauts were. Yet, the tech giants show remarkable resilience — despite it being easier than ever to build a Facebook or Google clone, the big guys seem to remain pretty competitive.

Why do some firms stay competitive for longer?

Firms obviously compete with the aim of making money, working under continuous competitive threat of other market players — the ocean out there is red from the competitors’ blood. Operating in such a competitive environment, what keeps some businesses competitive over long periods?

Flamholtz & Hua in their 2003 work define the intangible company infrastructure as resource management, operational systems, planning systems, and corporate culture. They then further posit that since markets can be penetrated by competitors and products can be copied, it has to therefore be that infrastructure that creates a long term strategic advantage.

Flamholtz and Hua’s research implies that it’s the interaction of culture, management and planning systems, and resource management that makes companies successful in the long run (as measured by achieving higher return on equity over longer periods). Flamholtz illustrates this as the pyramid of the organizational development, thereby indicating that culture is only the last step in building a company’s internal infrastructure.

Tactics vs Strategy

Majority of strategic decision making at companies is actually tactics with time horizons of at most a year or two. Sustained competitive advantage is by definition concerned with longer time horizons: more likely to last decades than quarters.

Since many things change in the competitive landscape over a decade, only the company infrastructure can remain stable, allowing the company to adapt to the changing environment and remain competitive.

Culture (and associated management systems) IS strategy.

Everything else is just quarter-to-quarter, year-to-year tactics.

What does this mean for managers?

Your team culture is replicated by people in your team copying behaviors which get rewarded. View your culture as your key strategic pillar. This means you should ask yourself what types of behaviors or values will make the company resilient or even antifragile in the long term.

In my line of work (managing software engineers), this includes cultural artifacts like career ladders, postmortems, design docs, demos, architecture review sessions, etc. We show what good looks like (i.e. what behaviors we want or don’t want) but also make space for learning and debating technical nuances with each other. This is critical if you want to avoid structural weaknesses such as Key Person Risk or your company becoming a one trick pony.

Solid, decentralized decision making is key to making your organization nimble and proactive over the long term. And you need to empower those decentralized decision makers by making room for them to learn the context around them — and allow them to make mistakes! That’s the main way your culture becomes self-replicating and your strategy adapts to its environment as a result.

Doing these things takes a lot of time and it’s easy to deprioritise them in favour of doing real, Q4-relevant work. But in the long term that’s what matters. In my opinion at least :)

(*) There is no evidence Peter Drucker ever said that, even though it does sound like something he would say.

Let’s first address the elephant in the room

Betting is still taboo in most of the world, so I had to do some introspection whether I wanted to be a part of that world. On one hand, most human activity has side effects (aka externalities) so it’s easy to dismiss gambling as being yet another business with negative externalities. On the other hand, I didn’t want to dismiss these externalities without exploring them deeper.

As with most values-based decisions, this is primarily a personal question so I don’t think my judgement can be universally applied, but my thought process might resonate with some, so I thought it worth sharing. I realized that I wouldn’t have qualms working in other industries with significant negative externalities such as the oil & gas or the auto industry, alcohol producers, the pharma industry, or the tech giants. Most societies today don’t do a great job of compensating the negative externalities of these industries, but the situation is improving — our cars are greener, alcoholism treatment and awareness is increasingly available, and even big tech is getting closer scrutiny these days.

Betting and gambling is heavily regulated and taxed, with increasingly strict responsible gambling rules and limits. So how big of a problem is problematic gambling? This Wikipedia article suggests roughly 0.5–1% of the population exhibits problematic gambling behaviour. This survey puts the number for Romania (Superbet’s home turf) around 0.6%. Alcoholics account for around 5% of the population according to this other Wikipedia article.

Given that preventing access to gambling for problematic players is not only doable, but mandated by law and increasingly strictly regulated, I decided I’m fine with making the call.

Betting Software Ecosystem is surprisingly complex

The first thing that overwhelmed me when I joined was the sheer complexity of sports betting, so thought it might be useful to shed some light on that. The basic idea of a betting business is one of mediation — in an ideal scenario the betting operator is just an intermediary between two punters who want to bet on the same event. The odds reflect demand/supply for each of the outcomes, minus the operator’s margin.

In practice, you don’t always live in an ideal scenario (e.g. the event is not prominent so supply or demand are not liquid enough), so providing a smooth betting experience regardless is harder than I expected. Also, the distribution of customer interest for betting events is a power law, making for very spiky customer traffic patterns.

The software ecosystem dealing with all this complexity can be organized in three major product areas:

• Trading side: this is the part of the ecosystem that ingests sports and betting market data and builds the betting offer through offer management and risk tooling. A 100+ people ops team (called traders because determining the odds is essentially a supply/demand problem) use these tools to make sure Superbet has the broadest, deepest, and best priced betting offer on the market.
  A Google Search equivalent of this would roughly be Indexing.
• Betting side: the transactional, customer-facing part of the system, that includes the betting clients apps and the systems enabling the betting ticket lifecycle.
  This roughly corresponds to Serving in Search.
• Operations side: running a multi-channel, multi-country business of this scale requires a lot of operations support tooling that enable us to operate the retail betting shop network, customer support issues, marketing and CRM ops, etc.

My experience so far

While our DAU or QPS metrics are nowhere near Google scale, the complexity of problems being solved is comparable, especially when you take into account that you have to do a lot of things from scratch, without relying on the massive Google infrastructure.

The main thing I loved about my new team was how easy it was to try out new things — ideas, tools, processes, or ways of working. While Google is probably reasonably fast for its size, this is just in a different ballpark. We talk about something, decide to try it, and start doing it next week.

While nimbleness of a smaller team is inspiring, some things just work much better at scale. Most Googlers don’t have to know how to configure or manage a Borg cell, distributed in-memory cache, or a large database system correctly. If you solve the quota issues, you can count on the system working without knowing a lot of internal SRE magic (modulo the occasional cell deprecation or maintenance ;-) ).

We mostly had to figure these things out on our own and learn way more details than we planned to. This doesn’t only apply to engineering practices. We also had to figure out how an engineering career ladder should look like, introduce design docs and postmortems, figure out oncall and escalation procedures, quota planning, etc.

My initial thinking was to just do it the Google way but we quickly learned that while most of Google’s ideas are generally applicable, the devil really is in the details, and there’s no one-size-fits-all approach that works across organizations of different sizes, different cultures, and industries.

So, what’s next?

It is a cliche, but things are just getting interesting. On the one hand, Superbet is expanding its operations geographically, providing a steady stream of localization, integration, and migration challenges. On the other, we have a ton of exciting product ideas we want to try out.

We see ourselves as primarily a tech company working in the betting industry, rather than a betting company. That puts us in a position where due to the heavy investment in our software ecosystem we can try out new product ideas and pivot as we learn more. This is an industry dominated by major players using off the shelf b2b solutions, so we think fully owning the whole customer experience is our long term competitive advantage.

Our next big organizational change is the pivot towards self-sufficient and empowered product teams. This will open a lot of interesting engineering challenges. For example, it’ll be crucial to figure out how to build platforms that are easy to contribute to, yet rock solid in terms of stability and performance. Product teams will also require a mindset shift; encouraging engineers to get actively involved in feature discovery, hypothesis testing, and delivery is radically different than what they’re used to.

We’re on a long learning curve. Feel free to reach out if it sounds interesting ;-)

I recently graduated from WBS with an MBA, and as part of that I did a dissertation in the field of financial analysis. The full dissertation is available for download here. Do get in touch if you have feedback, I’d love to hear your thoughts. What follows is a short and informal summary of this work.

Financial analysts’ short term predictions tend to be included in the share prices. A clear indication of that is how stock prices fall when earnings announcements miss forecasts. At the same time, a growing body of literature suggests that analysts tend to be too optimistic with their long term forecasts. In other words, the market trusts analysts in the short term, but ignores their forecasts for the long run. I explored this conundrum as part of my dissertation.

I first checked whether analysts were indeed too optimistic with their long term forecasts, and then looked at whether this optimistic bias might come from analysts preferring higher-growth industries. This was done by examining over 800 analyst reports covering two industries from the Dow Jones index (manufacturing and IT), over a span of ten years (2009–2019).

The main findings of the study are as follows:

• Median observed long term growth rate was 3.5% for analysts and 3% for the market in general. Analysts were indeed found to be more optimistic than the market.
• Evidence of a stronger optimistic bias towards a lower growth industry (manufacturing) rather than a higher growth industry (IT) was found.

The last finding was unexpected. One potential explanation for this could be that in the post-crisis period analysts preferred the safety of well known manufacturing powerhouses such as Boeing or Caterpillar. This hypothesis could not be confirmed with the collected data set. A follow up study including more firms and other time period should be done because some indications of outliers in the sample were found, and given the small sample size their impact is strong.

I have to thank Dr Shahed Imam for his guidance while working on this, it was good fun and I learned a lot. Looking forward to applying the things I learned in the MBA and this dissertation to put some skin in the game :)

Zahvaljujući vrlo visokoj razini transparentnosti izvještavanja (dosta naših državnih kompanija nisu toliko transparentne), naučio sam i više nego što sam očekivao. U nastavku slijedi kratka analiza financijsko-strateškog stanja Rimac Automobila i odgovor na pitanje je li RA stvarno “hrvatski Tesla Motors”, kako često znaju glasiti novinski naslovi. Na kraju ću probati glumiti strateškog konzultanta, tj. pametovati iz daljine.

Rimac Automobili su još uvijek mala firma

Unatoč jakom rastu prihoda (vidi graf niže), RA su još uvijek mala firma. Prema kriteriju prihoda, u Rimcu su tek 2018. došli do razine prihoda od 10 milijuna eura, duplo manje od npr, Pekare Dubravica koja je s tek stotinjak zaposlenih više imala duplo veći prihod.

Također, s 316 prosječno zaposlenih tijekom 2018., prosječni prihod po zaposlenom iznosi mršavih 200 tisuća kuna, jedva dovoljno za pokriti prosječnu bruto plaću zaposlenika (malo preko 190 tisuća godišnje) pa stoga ne čudi da je firma već nekoliko godina u gubicima.

Naravno, promatranje RA isključivo kroz prizmu prihoda i dobiti je besmisleno — riječ je firmi koja tek gradi vlastite kapacitete i stoga zahtjeva veliku količinu kapitala. Prema Damodaranovom modelu razvoja kompanije, RA su još uvijek u ranoj start-up fazi: razvijaju ključne organizacijske sposobnosti (znanja, alate, odnose s dobavljačima i kupcima) i još uvijek nemaju proizvod koji mogu profitabilno prodati u velikim količinama.

Ako su kao što je Mate tvrdio, uslužna R&D tvrtka ili proizvođač komponenti za veće igrače, nejasno je čemu distrakcije poput Greypa ili samih automobila. Ako pak rade aute, onda su sigurno tek u počecima jer je po svemu sudeći ukupan broj proizvedenih automobila jednoznamenkast. Trenutno se čini da pokušavaju raditi oboje: komponente jer su vjerojatno trenutno profitabilne, ali su dugoročno commodity što ih čini strateški manje atraktivnima, i aute na kojima grade brand i trude se profilirati se u high-performance niši poput npr Koenigsegga.

Što (i kome) prodaju?

2015 je bila prva godina u kojoj je RA imao ozbiljnije prihode (preko 20M kuna), a ujedno i prva godina kad su povećali transparentnost izvještavanja pa su i dostupni podaci o strukturi prodaje tek od 2015, ilustrirani na grafu niže.

Dvije stvari iskaču: prvo, prema očekivanjima, RA stvarno žive od usluga i drugo, puno manje očekivano, 2015. je prodaja usluga u Hrvatskoj bila većina prihoda — preko 23M kuna. Bilo bi zanimljivo vidjeti tko u Hrvatskoj ima potrebe za uslugama RA. Što se tiče broja prodanih Concepta One, ti podaci nisu eksplicitno navedeni u izvještajima ali se da naslutiti sljedeće:

• Izvještaj uz 2014. spominje potpisivanje ugovora o prodaji s kupcem i uplati avansa u iznosu 5.5M kuna. Ovo je vjerojatno prvi prodani Concept_One.
• Primici od prodaje proizvoda u 2017. su iznosili malo preko 24M kuna, što bi dalo naslutiti du su te godine prodana tri auta. Ovo je dodatno potvrđeno u bilješki za 2016 kod objašnjenja tehnikalija promjene knjigovodstvene politike.
• U 2017 su imali 7.4M kuna primitka od osiguranja, vjerojatno za auto koji je Hammond razbio u Švicarskoj, u skladu s cijenom od 1M EUR koja se spominjala u medijima
• Bilanca za 2018. Ima 18M kuna “proizvodnje u tijeku”. Kako je Concept_Two tek u razvoju, za vjerovati je da on još uvijek ne čini većinu te stavke — većina razvojnih troškova se knjiži u nematerijalnu imovinu. Stoga, moguće je da se iza ove stavke kriju još barem dva ili tri Concept_Onea u završnim fazama proizvodnje. S tim bi svih osam Concept_Onea koje je Mate najavio proizvesti trebala biti na broju.

S druge strane, u periodu 2015–2018 se omjer potraživanja od kupaca i prodaje kreće u rasponu 49%-60%, indicirajući moguće probleme s naplatom. Budući da se većina potraživanja odnosi na kupce u inozemstvu, za pretpostaviti je da zvučna imena auto industrije s kojima RA surađuju dobro koriste svoju pregovaračku moć i maksimalno razvlače rokove plaćanja, dovodeći RA u situaciju da im netto radni kapital iznosi gotovo 30% “opipljive” imovine u 2018.

Kako se financiraju?

Kao što je već na početku članka navedeno, RA još nisu profitabilni a i za očekivati je da niti neće biti još barem nekoliko godina. Stoga je jasno zašto je skupljanje investicija jedna od glavnih Matinih zadaća, a ako je suditi po nedavnom intervjuu, i jedna od najstresnijih. Graf niže opisuje kako su se izvori financiranja RA mijenjali kroz godine:

U periodu 2011–2014, većina financiranja je dolazila od posuđivanja od banaka, osnivača, sestrinskih firmi i slično. To je ujedno i period iz kojeg vjerojatno potječe priča “tata mu financira firmu” koja kola po raznim komentarima naših portala. Iako se dug kao izvor financiranja u ovom dijelu životnog vijeka kompanije smatra lošom praksom jer organičava mogućnosti reinvestiranja i podiže rizik (za razliku od dioničara, vjerovnici očekuju svoje novce natrag — Damodaran je iz sličnih razloga kritizirao Teslu nedavno), Mate vjerojatno nije imao drugog izbora. Za razliku od Kalifornije, rizičnog kapitala u Hrvatskoj nema puno, pa je koristio izvore koje je mogao. Čim su RA “došli na radar” drugim igračima u autoindustriji, kapitalna struktura kompanije se okrenula u korist dioničara. Iako je takva struktura u skladu sa stupnjem razvoja firme, sa sobom nosi i moguće strateške izazove radi pozadine novih dionicara.

Zaključak

Kratka verzija:

1. Čini se da su proizveli svih osam Concept_Onea, i prodali barem 5.
2. Obitelj Rimac je uglavnom financirala firmu do 2014, ali od onda to uglavnom nije istina.
3. Uglavnom se bave uslugama razvoja za druge, ali čini se da to žele promjeniti. Nije sigurno da će im to uspjeti.

Duga verzija:
Prerano je govoriti o Rimac Automobilima kao o hrvatskoj Tesli. Po svim važnim parametrima (stupanj razvoja proizvoda, proizvodni volumeni, kapitalna struktura, razina prihoda), Tesla je tinejdžer dok su RA još uvijek u pelenama. Osim toga, radi Muskovog backgrounda u Tesli nisu morali raditi razvoj za druge (u manjem obimu su radili za Toyotu, ali i onda je to bio manji dio prihoda od prodaje Roadstera), pa i ne čudi da kompanije imaju različiti razvojni put.

Ipak, pošteno je reći da su RA profitirali od Tesle — čini se vjerojatno da je nedavni val investicija (Porsche, Hyundai) i suradnji (Aston Martin, Jaguar, Seat, Pininfarina, Koenigsegg) iz industrije došao upravo radi tektonskog poremećaja u industriji koji je Tesla uzrokovala s uspjehom svojih modela S, 3, X — tradicionalne firme se osjećaju ugroženima, pa vide RA kao brži put za dostići konkurenciju od vlastitog razvoja ab ovo.

Dugoročno, ako se cijela auto industrija elektrificira (što se sad već čini izglednim), komponente i razvojne usluge koje RA nudi svojim klijentima su ključne kompetencije firmama u industriji, pa se ne čini logičnim imati ta znanja izvan firme. Iz te perspektive se manjinski udjeli u Rimcu mogu činiti kao strateški potez auto firmi za buduće preuzimanje (kao svojevrsna call opcija u slučaju bankrota RA).

Partnerstvo s Camelom s druge strane se čini kao jednostavan put do skaliranja proizvodnje komponenti za RA, i tehničkog unapređenja proizvoda za Camel koji još uvijek uglavnom prodaje klasične olovne akumulatore. Glavni izazov u tom partnerstvu se čini asimetričnost — jednom kad Camel nauči raditi moderne baterijske sustave od Rimca, što ih spriječava izaći iz partnerstva u državi poput Kine s povijesti labave interpretacije intelektualnog vlasništva?

Na kraju, proizvodnja vlastitih hiperauta se čini kao logični odgovor na obje strateške dileme, ali dolazi sa ogromnim zahtjevima za kapitalom. Čini se da im je trebalo oko pet godina i 40 milijuna kuna (gledajući kapitalizirane troškove razvoja) za produkcijsku verziju Concept_Onea. Za auto poput Concept_Two koji se očekuje prodavati u većoj seriji, i tehnički je još zahtjevniji, te bi brojke lako mogle biti i 5–6 puta veće. Iz te je perspektive jasno zašto RA igra obje karte — uz prodaju razvoja i komponenti, razvijaju i aute.

Naravno, budućnost RA u ovom obliku još uvijek nije sigurna jer su još uvijek daleko od pozitivnog cashflowa. S druge strane, razina investicija od investitora iz industrije daje nade da će elektro-auto industrija u nekom obliku u Svetoj Nedelji preživjeti — čini se logično da će i u slučaju bankrota RA trenutni stakeholderi htjeti iskoristiti razvijena znanja i tehnologije, jer bi razvoj takvih znanja od početka trajao godinama.

There are few companies in the US that represent the classic rags-to-riches American dream as much as Ford does. I previously wrote about their early differentiation failure, but this time I wanted to explore how they are doing today. Since Ford is a global company operating in multiple segments, the focus of this essay is a strategic analysis of Ford Motor Company in the New Light Vehicle US market (passenger cars, SUVs, light trucks).

The industry is brutal and bound to get worse

Minimum Efficient Scale (MES) of production is very high, leading to high capital requirements. Smallest mainstream producers sell just under 300,000 units per year, in line with Huang’s estimate of MES for automotive manufacturing at 250,000–300,000 units (link). This makes new entrants rare and typically narrowly focused on niches. It also explains why Tesla only became profitable after reaching ~300k annual production rate.

Switching costs are close to zero and cars cost a lot of money, making buyers price sensitive and often delaying purchases in recessions. Car manufacturers successfully defend against this with branding — roughly half the population consider themselves loyal to car brands. Still, the industry is dominated by consumer tastes — Henry Ford’s famous quote on Model T color options simply wouldn’t work today. This is well illustrated by the dramatic difference in consumer tastes between Europe and the US or the fact that ~15% of Americans considered buying an electric vehicle in 2018, up from ~0 only a decade ago.

Another force increasing market pressure is market saturation — vehicle ownership and distance-driven rates in the US appear to have peaked around 2006.

Faced with a stagnating market, driven by price sensitive customers, with no opportunities for collusion due to limited market power and regulatory scrutiny, car manufacturers’ returns on invested capital are unsurprisingly low: PwC’s 2017 research indicates only 5% industry-wide return on capital, below their cost of capital.

An industry-wide shift towards electric vehicles is going to require a massive capital investment to maintain the minimum efficient scale and build out the necessary charging infrastructure. Heavily indebted companies will not be well positioned to do this and face an existential threat as a result.

How is Ford positioned to deal with this?

The following diagram illustrates industry strategic groups (see this spreadsheet for details on sources). Manufacturers are plotted against two dimensions: average selling price, and the number of brands. Number of manufacturer’s brands is used to estimate whether the manufacturer is pursuing a broad strategy or a narrow one. Size of the bubble represents market share.

FCA, GM, and Honda appear to be pursuing cost leadership strategies where competitive rivalry is typically strongest, due to price being a key decision factor. Ford seems to have avoided that battle and we next explore how.

Out of 2.58M vehicles Ford sold in the US in 2017, 1.1M were pickup trucks (see annual report for details). This is an excellent segment to differentiate in: trucks have a higher average selling price, lower emission requirements, and consumers favour US brands in the truck segment — over 85% of truck sales are by “big three” manufacturers. Additionally, consumers are increasingly shifting towards trucks — sales volumes overtook cars in 2012. Ford is a top player in this lucrative segment, a position strengthened by their $2.54B advertising spend in 2017, second only to GM. If we consider that GM has a broad brand strategy, Ford is probably the biggest automotive spender in advertising per brand.

To conclude, Ford is well positioned for current market conditions. We next consider key drivers of change in the industry and propose ways Ford could adapt.

Direct Selling

To capture more value, Ford should consider integrating their dealerships and sell directly to consumers. Franchising is required by law in most US states, but as Tesla showed, it can be done. Other manufacturers are likely to follow, putting Ford at a serious competitive disadvantage if they don’t.

Autonomous Driving

With Waymo announcing start of operations for their self-driving taxi service, and with GM’s Cruise subsidiary announcing 2019 launch, autonomous driving appears at an inflection point: self-driving taxi services could be common within a decade.

As a result, taxi prices will drop substantially, substituting car ownership for a significant fraction of the urban population. New vehicles demand will further drop but also structurally change the competitive environment: automated taxi services will dominate car purchases, increasing their bargaining power. When they become a significant fraction of Ford’s sales, autonomous taxi companies are likely to try acquiring car manufacturers to vertically integrate. Similar to buses, end-consumers will care less what car brand they’re driven in, killing Ford’s brand “moat”. Deloitte calls this scenario “Hardware Platform Provider” in their 2017 study, reminiscent of PC manufacturer’s fate in the ‘90s.

Ford should therefore invest heavily in developing their own autonomous mobility service. They are working on it but doing it themselves might be challenging. If their lag behind the competition proves too big, they might consider investing in one of the existing players with which they aren’t direct competitors. For example, FCA recently announced a partnership with Intel and BMW.

Electrification

A key technological challenge in electric vehicles production is energy storage and charging capabilities. Battery production capacity needs to dramatically increase to meet the demands of the growing electric vehicle fleet.

A key strategic decision is whether to own battery production or to source externally. This depends on how will the battery supplier market evolve: will it consolidate, giving few key suppliers great bargaining power over Ford, or will all car manufacturers outsource battery production, thereby creating a large commoditized market where battery suppliers will not have a lot of bargaining power.

Only Tesla is currently pursuing in-house battery production: VW, GM, Honda, and FCA decided to source externally. Sourcing externally is a better choice in the short to medium term — it is less capital intensive, and most competitors are doing it as well, creating limited competitive advantage opportunities in battery production. Given Ford’s high leverage, they probably can’t afford doing it themselves, so a long term battery supply contract might be within the realm of feasible.

Another key aspect to consider is what are the “petrol stations” of the future going to be like. Should manufacturers build their own networks, or outsource that activity to others? Ford already struck a deal with Daimler, VW, and BMW to jointly build the charging network in Europe, but there is no news of a similar deal being struck in the US. GM outsourced their US charging network to evgo.

Similarly to battery manufacturing, Ford probably can’t afford the massive capital investment this requires, and pairing up with companies that aren’t their direct competitors like Daimler, VW, or BMW makes sense. Another candidate to partner up with could be oil companies — re-purposing petrol stations is less demanding than starting from scratch and oil companies have similar incentives to do this as Ford — survival.

To conclude, I think Elon might be right — things don’t look good for Ford. While they’re currently doing ok, they don’t appear well prepared for the incoming industry changes. A recently announced VW alliance might be a step in the right direction to deal with these challenges (leverage your ally’s resources rather than building your own), though it carries its own risks by becoming too dependent on VW.

Ford was originally published in HackerNoon.com on Medium, where people are continuing the conversation by highlighting and responding to this story.

The Real Frank Zappa Book, 1999, by Frank Zappa with Peter Occhiogrosso

Prema Zappinom kriteriju, Hrvatska bi uskoro mogla imati ugrožen status Prave Zemlje. U ostatku teksta ću pokazati kako je Croatia trenutno na putanji za propast, uglavnom kao rezultat lošeg upravljanja u nepovoljnim tržišnim uvjetima.

Svi podaci izneseni su izvučeni iz godišnjih financijskih izvješća Croatia Airlinesa i obrađeni za potrebe teksta u Google spreadsheetu.

Avioprijevoz je zahtjevna industrija

Avioprijevoz je jako zahtjevna industrija za natjecanje. Warren Buffett je 2007 napisao da bi investitori u aviokompanije, da se mogu vratiti u prošlost, vjerojatno upucali Orvillea Wrighta usred leta u Kitty Hawku.

Industrija je zahtjevna jer ključni dobavljači (proizvođači aviona, motora i aerodromi) imaju de-facto monopol pa su troškovi visoki, a proizvodi slabo diferencirani pa se za većinu karata prijevoznici natječu samo cijenom. S druge strane, EU regulativa i EU-US open skies agreement čine tržišnu utakmicu gotovo slobodnom. Zato letimo sve jeftinije dok vlasnici aviokompanija “jedva krpaju kraj s krajem”.

Zato ne čudi što je avioindustrija u EU prošla kroz nekoliko rundi “brakova iz nužde” u zadnjih desetak godina: Lufthansa je kupila Austrian i Swiss iz bankrota, British Airways i Iberia su se udružili, kao i KLM i Air France. Cilj udruživanja je uglavnom bio postizanje boljih ekonomija razmjera i smanjenje tržišnog pritiska kroz manji broj tržišnih igrača.

Jednom takvom braku iz nužde se nada i naša vlada kao rješenju za probleme Croatia Airlinesa.

Hrvatsko tržište je izrazito sezonalno

Kao što dijagram ispod ilustrira, potražnja je na hrvatskom tržištu izrazito sezonalna — u najslabijem mjesecu kroz hrvatske zračne luke prođe gotovo deset puta manje putnika nego u najboljem.

Prema podacima EU, po ovom je kriteriju Hrvatska na vrhu EU — u većini ostalih EU država taj omjer je u rasponu 2–3. Nakon Hrvatske, najgora je Grčka s faktorom 5. Sezonalno tržište predstavlja kompanijama izazov u operacijama jer im kapaciteti u ostatku godine zjape prazni. Kao rezultat, Croatia je profitabilna samo šest mjeseci godišnje.

Konkurencija je jaka i Croatia polako gubi

Kao što se vidi iz prethodnog dijagrama, u periodima visoke potražnje za HR letovima, Croatia ima veliku konkurenciju ostalih europskih avioprijevoznika. Ovo se vidi i na Croatijinim financijskim rezultatima: u periodu 2006–2017 je prosječna cijena leta je pala 16% (853 => 722 kn). U istom periodu su operativni troškovi po prodanoj karti skočili 3.5% (564 =>584 kn).

Ovo nas vodi do idućeg zaključka.

Croatia je neučinkovito vođena kompanija

Zahvaljujući malom broju letova i putnika, CA ne uspijeva postići efekte ekonomije razmjera. Na primjer, imaju 8 ureda van Hrvatske u kojima radi 38 zaposlenika. Zaposlenici ureda u Zurichu imaju posla sa samo dva leta dnevno. Vjerojatno je da bi isti ljudi mogli odrađivati više letova.

Ovo se pak najbolje vidi u omjeru putnik-kilometar/zaposlenom — Croatia je u 2016 odradila 1.53 milijuna putnik-kilometara po zaposlenom, dok je u Lufthansi taj broj iznosio 1.84. RyanAir je na 9.88. Croatia se u 2017 približila Lufthansi s 1.75, najvjerojatnije kao rezultat korištenja dva wet lease aviona u sezoni.

Struktura zaposlenika u CA je previše top-heavy. Operativno osoblje (uključuje ground/cabin/cockpit crew, tehniku, prodaju) čini samo 65.6% zaposlenih. U RyanAiru je taj omjer na 94.8%. Budući da je CA državna tvrtka, lako je povjerovati da postoji i pristojna količina “uhljeba”.

Passenger load faktor (omjer iskorištenih i dostupnih sjedalo-kilometara) je konzistentno ispod prosjeka industrije. U 2017 je postignut najbolji load factor u povijesti Croatije 74.4% (prosječno tek 67% 2006–2017). Prosjek avioprijevoznika u Europi je bio 81.5% dok RyanAir, lider industrije, stoji na 93%.

Čini se da Croatia ima neadekvatnu flotu. Airbusi su sigurno prekapacitirani avioni van turističke sezone, ali postoje naznake da to vrijedi čak i u sezoni. U godišnjem izvješću za 2013 se uprava hvali kako su iznajmili A320 s flotom u Tunis u razdoblju 31.5. — 1.9. Izgleda da je wet lease bio profitabilnija opcija od korištenja vlastitog aviona s posadom u sezoni.

Svi ovi faktori su doprinijeli tome da je čak i nakon provedenog restrukturiranja 2012, Croatia samo naizgled profitabilna kompanija. Taj privid se održava nizom trikova kojima se imovina kompanije pretvara u buduće obveze:

• Najzvučniji takav primjer je bila prodaja LHR slotova 2017 za 138 milijuna kuna. Bez tog prihoda bi Croatia bila oko 100 milijuna kuna u gubitku, umjesto iskazane dobiti od 39.8 milijuna kuna.
• 2015 su napravili “sale-and-lease back” motora za Airbus. Nemamo podatke za koji iznos je motor prodan, ali znamo da je Croatia te godine bila u plusu gotovo 15 milijuna kuna, dok je prihod od prodaje dugotrajne imovine iznosio skoro 30 milijuna kuna (izvješće za 2015, p85).
• 2014 su napravili “sale-and-lease back” Airbusa 9A-CTJ i tri motora za Airbuse. Opet nemamo točne podatke o prihodima koji su nastali radi toga, ali znamo da je prihod od prodaje dugotrajne imovine iznosio oko 60 milijuna kuna (izvješće za 2014, p85). Croatia je te godine iskazala dobit od 8 milijuna kuna.

Na dobar dio ovih odluka je Croatia bila primorana od strane Europske komisije, kao cijena za vladino opraštanje dugova (na što se s pravom gleda kao oblik državne subvencije) u sklopu restrukturiranja iz 2012. Croatiji će uskoro nestati dugotrajne imovine za prodaju, pa je za očekivati povratak k negativnim financijskim rezultatima, naročito u slučaju neke nove recesije.

Zaključak

Croatia posluje na tržištu s izraženo sezonalnom potražnjom s flotom koja je uglavnom fiksna i optimirana za tri mjeseca vršne potražnje. Zbog male veličine tržišta ne uspijeva postići ekonomije razmjera u poslovanju koje je ionako opterećeno prevelikim brojem neoperativnog osoblja. To rezultira visokim jediničnim troškovima koji u kombinaciji s niskim load factorima čine većinu mjeseci u godini neprofitabilnim za Croatiu.

Wet leasing aviona u vrhu sezone je malo popravio situaciju, međutim Croatia se iznutra nije promijenila. Vlada je očito svjesna strukturnih problema u kompaniji koji se jedino mogu riješiti politički nepopularnim metodama poput otpuštanja i smanjenja vlastite flote.

Vlada stoga već godinama priča o prodaji kompanije “strateškom partneru”, međutim ponude ne dolaze. S obzirom na poziciju Croatije i stanje tržišta, to i ne čudi. Vjerojatno čekaju da Croatia propadne, pa jednostavno preuzmu njene linije, bez preuzimanja njenih obveza.

Brak iz nužde kojem se vlada nada ne djeluje izgledno.

A quick reminder of IBM’s glorious history

IBM is an old company by any standard, and in computer history terms, it’s a bona-fide dinosaur — they existed before the first electronic computers. In fact, it is so old that there is a very long wikipedia article dedicated to its history. To name a few notable achievements in company’s history, before giving us the PC in 1981, IBM practically invented mainframes, helped land humans on the moon, controlled the Space Shuttle, invented relational databases, the first optimizing compiler, and launched the first commercially available laser printer (which to be fair, was invented by another IT industry dinosaur — Xerox).

IBM’s business transformations

IBM had to keep reinventing itself to avoid the fate of dinosaurs as technology kept changing, and it’s been quite successful at that. They successfully transitioned from mainframes to personal computers in the 80’s, pivoted out of commodity hardware (disposed of the printer business in 1991, personal computer business in 2004, x86 servers in 2014) into business services (boosted by the 2002 acquisition of a part of PwC Consulting). This transition is illustrated by the following chart:

However, they seem to have missed two major opportunities in the IT world at the beginning of the 21st century — cloud and mobile. To address that, IBM launched new “strategic imperatives” in 2012, that include analytics, cloud, cognitive (IBM’s generalization of the term “AI”), social & mobile, and security.

IBM joined the cloud game through the SoftLayer acquisition in 2013, and on the mobile front tried to get a piece of the “business apps on iOS” market, but not to much success. The 2016 annual report merely mentions the mobile business as growing, but no specific numbers are given (accounting rules require separate disclosure of any substantial business segment which implies mobile is not a substantial business). Cloud is reported as bringing in $10.2 billion in 2016, but it should be noted that roughly half of it ($5.6B) comes from “foundational offerings — where the company provided software, hardware and services for clients to build their own clouds.” (IBM 2016 Annual Report, p57). They’re essentially building clients’ data centers and count that towards their cloud revenue.

On the artificial intelligence (AI) front, IBM was an early player in the market — they famously beat Kasparov at chess in 1997, and won Jeopardy in 2011. From then on IBM heavily invested in turning Watson into a business-friendly AI platform by investing in Watson Health (medical decision support), Watson IoT (Internet of things big data analytics platform), and Watson Platform (a data analytics platform). According to the 2016 annual report, Health is the strongest focus in this area with IBM employing over 7,000 people on Watson Health.

While it has been rolled out to dozens of hospitals worldwide as of end-2016, it’s success does not seem certain — one would expect skyrocketing revenue if the solution outperformed doctors in a significant way.

Is IBM still a wonderful performer?

After going through a decade of slowly growing revenues in the early ’00s, IBM posted revenue loss for 22 consecutive quarters, returning 2016 annual revenue slightly below the 1998 level. The management seems to have managed oscillating revenues well as earnings nearly tripled from $7,524 million in 2002 to $22,540 million in 2012. This was probably one of the reasons why Warren Buffett, well known for avoiding tech stocks, invested in IBM back in 2011.

Why are revenues falling?

IBM’s two main revenue streams, sales and services are falling since 2011 (although services seem to have slightly recovered in 2016). This can be explained by the changing IT landscape in the past seven years: companies are rapidly moving their data centers to the cloud (which isn’t powered by IBM hardware), and require more IT consulting in areas (like user experience design, mobile apps, and digital marketing) where IBM doesn’t have a significant skill set.

IBM recognized this and reacted by building it’s “digital economy” know-how by starting it’s own digital agency: IBM Interactive Experience in 2014, which it boosted by a series of acquisitions — from 2015 until 2017 they acquired six consultancies. That could be what Buffet was referring to when he said that “IBM is a big strong company, but they’ve got big strong competitors, too” after selling roughly a third of his stake in 2017.

Is Cuban right?

Cuban’s bashing of IBM can likely be summarized with the following excerpt from the 2013 annual report

In the 14 years from 2000 until the end of 2013, IBM generated roughly $165 billion of cash flow, 81% of which was returned to shareholders through stock buybacks and dividends. Add to that $32 billion of acquisitions in the period, and you get an understanding of what Cuban was referring to — IBM has been reacting to the market changes (as opposed to changing the market as they used to) by buying market players in areas where they are weak.

Regarding stock buybacks, they had a proclaimed goal of hitting a specific earnings per share figure by 2015, so one can see how Cuban thinks that IBM’s management is focusing too much on their shareholders instead of focusing on their customers and competitors. That could explain how IBM’s management missed the mobile, cloud, and AI opportunities described above.

Looking at revenue per employee performance metric, IBM lags in revenue per employee against its competitors like HPE, Amazon, and Microsoft, although they’re still ahead of pure services companies like e.g. Accenture. This explains why they want to move significant parts of their workforce oversea to cut costs — they expect the trend to continue.

Conclusion

As stated before, in the past decade IBM’s management focused too much on their investors instead of on company growth. As a result, they missed a lot of opportunities since 2005:

• Who else could have sold the cloud to enterprises (while it was still considered “unsafe”) if not IBM?
• Similarly, who else could have sold mobile apps, design thinking, and all modern “digital trends” to big enterprises if not IBM?

Still, not everything is gloomy. The services business seems to be half way through the “digital transformation” (I know, it sounds strange to think of IBM as needing digital transformation, but they do. Maybe “modernizing” is a better word?), their services backlog is still very healthy (more than two years worth of at the 2016 pace), and mainframes are apparently still a thing (most of their Q4 2017 growth actually came from the mainframes). So, IBM is likely here to stay. And despite what Mark Cuban says, they are still at tech company.

There are even a few “moonshot” opportunities for growth in IBM:

• Watson could still turn out a success. After more than three years “in production” that does not seem likely though — Google seems to be the key player in AI, with AlphaGo, TensorFlow, and Cloud AI APIs. They also started experimenting in the health sector.
• Blockchain opportunities through their Hyperledger project (for example, they recently signed the Maersk deal)
• They still seem to be a key player in Quantum Computing

There is a big if around all items on this list, so growth for IBM, even in the long term does not seem like something worth paying for a lot.

IBM: Big Blues was originally published in HackerNoon.com on Medium, where people are continuing the conversation by highlighting and responding to this story.

The hypothesis I’ll present here is that Edsel’s failure is a prime example of the weaknesses in the Waterfall product development model.

Why Did Ford Want to Start a New Division?

Despite having mid-range and high-end makes in Mercury and Lincoln respectively, Ford was mostly operating in the lower end of the market in the early 1950’s. A study conducted in 1952 concluded that over 75% of Ford’s production was in the economy range of the market (Dicke, 2010, p487). This was a great position to be in during and after the Great Depression, but with the rising GDP levels after WW2, a significant fraction of the population was expected to move into the middle-income range.

The growing middle class meant more people will be buying cars in the middle market segment soon and Ford had proof that Ford owners traded up to other makes, mostly made by GM — as Lewis Crusoe, VP at Ford in the fifties said “We have been growing customers for General Motors”. (Brooks, 2014, p.32). Ford’s only mid-range make, Mercury was perceived as a muscle car brand not in the mid-range segment (Brooks, 2014, p.41).

What Went Wrong?

Ford started considering the Edsel in late 1952, and it took them almost two years to figure out what car they want to build, and another two and a half years to actually build it. That means there was a five year delay from the moment planning started until the release day which was internally known as “E-Day”. During that time, the US economy went through two recessions (1953 and 1958), which had a drastic effect on the car market, but also on user needs, wants, and demands.

Consumer preferences shifted towards smaller and more economic cars — the so called “compact segment”. The only such car on the market at the moment of Edsel’s introduction was Rambler, which sold remarkable 242,734 cars in 1957–1959, compared to only 118,287 Edsels in 1958–1959. In fact, at the end of the 1950s, all of Edsels competitors saw disappointing performance. Chrysler even killed DeSoto, Edsel’s direct competitor. The only Edsel competitor that did well at the end of the decade was Dodge, whose success mostly came from sales of Dart, Dodge’s entry in the economy segment.

Did Ford Learn Anything?

They did. The 1960’s Edsel lineup already had a compact car planned, but when Edsel was killed, they decided to sell the car as Comet and transferred it to the Mercury brand in 1962.

They also seem to have changed their approach to developing and launching new products, as exemplified by Ford Mustang, one of their most successful products of all time. While building the Mustang, instead of the “big-bang launch” strategy, Ford employed a more agile approach, resembling Lean Startup’s build-learn-measure cycle.

Mustang began its life as a concept car (Mustang I — their Minimum Viable Product) that was first shown to the public in 1962 at the US Grand Prix and went on to be shown at trade shows. Ford even toured the colleges to gauge their target audience’s interest. Based on that feedback, they did 13 different variants and eventually settled for Mustang II in 1963. They did another tour of showings and eventually launched the production version in 1964. Mustang was built in only 18 months. They got as close to “release early, release often” as possible with building cars in the sixties.

Conclusion

Mustang’s development was a complete contrast to Edsel’s two years long marketing campaign, followed by a big-bang launch, and ultimate failure. With Edsel they took a classic waterfall-ish approach to product development, which at first sight makes sense — waterfall is considered appropriate is when requirements are fixed and the product is well understood.

Ford certainly knew how to build cars at scale, and the requirements they set up in 1953 were based on sound reasoning and research. The issue was that they weren’t learning while they were building the car, and as a result, their initial assumptions about the car market were no longer true by the time the car launched. Luckily for them, they took a different approach with the Mustang, and it paid off handsomely.

You can find a similar take on the Edsel case, from an “innovation cycle in organizations” perspective here.

References

• Brooks, J: Business Adventures: Twelve Classic Tales from the World of Wall Street, Open Road Integrated Media, New York, 2014
• Dicke, T: The Edsel: Forty Years as a Symbol of Failure, The Journal of Popular Culture, Vol. 43, №3, 2010, pp.486–502.
• Wilder, S: Mini Case Study, Ford Edsel, University of Amsterdam, 2014
• https://en.wikipedia.org/wiki/American_automobile_industry_in_the_1950s
• https://en.wikipedia.org/wiki/Dodge_Dart#1960
• https://en.wikipedia.org/wiki/Donald_N._Frey
• https://en.wikipedia.org/wiki/Edsel
• https://en.wikipedia.org/wiki/Ford_Mustang
• https://en.wikipedia.org/wiki/Rambler_American
• https://en.wikipedia.org/wiki/U.S._Automobile_Production_Figures
• https://en.wikipedia.org/wiki/Waterfall_model

Edsel’s Waterfall was originally published in HackerNoon.com on Medium, where people are continuing the conversation by highlighting and responding to this story.

Uvod

Ideja ovog posta je pokušati vidjeti je li se ovakav razvoj situacije u Agrokoru mogao predvidjeti i je li Mercator potopio Agrokor.

Iako se vole prikazati kao vertikalno integrirana firma koja posluje u regiji (“od polja do stola”), Agrokor je u stvarnosti maloprodajna firma koja je do nedavno uglavnom poslovala u Hrvatskoj. Tek je nakon preuzimanja Mercatora Hrvatska prestala biti izvor većine prihoda, a i danas maloprodaja čini 87% prihoda grupe. Stoga je usporedba Agrokora s hrvatskim maloprodajnim tržištem relevantna.

Analiza je temeljena na javno dostupnim izvješćima Agrokora. Iako su nedavno iskazane sumnje u vjerodostojnost tih izvješća (na službenoj stranici Agrokor tvrdi da je pokrenut postupak revizije izvješća), to su najbolje trenutno dostupne informacije, pa je analiza i napravljena s njima.

Što znamo?

Agrokor je gubio tržišnu utakmicu

Kao što je prikazano na slici 1, iako promet od trgovine na malo u Hrvatskoj kontinuirano raste od 2013, Agrokorovi prihodi u istom periodu kontinuirano padaju (ako se izuzmu Mercatorovi prihodi u periodu 2014–2016).

Ovo se vidi i na izvještaju Agencije za zaštitu tržišnog natjecanja koja u 2012. procjenjuje Agrokorov udio na tržištu na “25–30%”, dok isti udio u 2015. procjenjuje na “20–30%”, unatoč činjenici da brojka za 2015 uključuje i Mercatorov prihod koji je 2012 bio procjenjen na “5–10%”. AZTN procjenjuje da je još 2014. zbirni tržišni udjel Mercatora i Agrokora bio “30–40%”.

Agrokor je trošio skoro svu zaradu na plaćanje kamata…

Kao što se vidi iz slike 2, pokrivenost troška kamata zaradom je u razdoblju 2011–2015 bila blizu 100% što znači da je Agrokor gotovo svu dobit trošio na plaćanje kamata. Slijedom pada prihoda i zarade, izgledno je da je krajem 2016. ta vrijednost pala ispod jedinične što znači da Agrokor jednostavno više nije pokrivao troškove kamata.

.. a za to vrijeme je gomilao dugoročni dug

Gledajući razdoblje prije preuzimanja Mercatora, Agrokor je gomilao dugoročni dug koji nije trošen na rast dugotrajne imovine nego na financiranje kratkotrajne imovine. Kao što je prikazano na slici 3, u razdoblju 2009–2013, dugoročni dug je porastao za preko 6 milijardi kuna dok je dugotrajna imovina porasla tek za 3.5 milijarde.

Ovo daje naslutiti da je novac dobiven po skupim kamatama utrošen na financiranje redovnog poslovanja, umjesto investicija u povećanje produktivnosti kapitala i zaposlenika. Ilustracija toga su tri obveznice izdane 2012. godine. Ukupno je izdano obveznica u vrijednosti 6 milijardi kuna po kamatnoj stopi od gotovo 10%, od čega je 4 milijarde utrošeno na refinanciranje postojećih dugovanja.

Tablica 1 daje ideju kako je ostalih 2 milijarde utrošeno. 170 milijuna kuna na preuzimanje Roto Dinamica a ostatak na financiranje redovnih poslovnih aktivnosti poput financiranja obrtnog kapitala, financiranja potraživanja i slično. Deset posto je jako skupa kamata za financiranje tekućeg poslovanja, naročito u niskoprofitnom biznisu kao što je maloprodaja.

Što se tiče preuzimanja Mercatora, ono je financirano van Agrokorove bilance jer ona to jednostavno ne bi mogla podnijeti. Omjer dužničkog i dioničarskog kapitala je narastao s 1.39 u 2009 na 2.94 u 2013 (vidi sliku 4). Skok nakon preuzimanja Mercatora je isključivo uzrokovan preuzimanjem Mercatorovih dugovanja. Dug za preuzimanje Mercatora je izvan ovog omjera.

Interesantan detalj vezan uz preuzimanje Mercatora se odnosi na naknadno knjiženje vrijednosti branda Mercatora u Agrokorovom godišnjem izvješću iz 2015 — prema njemu je vrijednost Mercatorovog branda 1.58 milijardi kuna. Bez te (iako opravdane prema IFRS) računovodstvene egzibicije je ukupna vrijednost Agrokorove imovine u 2015 zapravo pala 300 milijuna kuna, otprilike iznos za koliko je Agrokor prikazao rast imovine dioničara. Bez te promjene je omjer dužničkog i dioničarskog kapitala u 2015 bliže razini iz 2014.

Kako je onda Agrokor kupio Mercator?

Ivica Todorić se 2014. zadužio putem PIK-Toggle kredita za 475 milijuna EURa sa sljedećim namjenama:

• Do 120 milijuna EURa za otkup manjinskih udjela u Agrokoru (Todorić je 2014. držao 94% Agrokora)
• Ostatak (355 milijuna EUR) za namjene: poslovanja Agrokora, potencijalno preuzimanje Mercatora, isplata dividende Ivici Todoriću

Od toga znamo da se dogodilo sljedeće:

• Ivica Todorić je 2014. dokapitalizirao Agrokor s 1.91 milijardu kuna (otprilike 255 milijuna EURa). S tim novcem je kupljen Mercator. Prema Agrokorovom izvješću za 2014., akvizicija Mercatora je Agrokor koštala 1.96 milijardi kuna (oko 260M EUR).
• Prema istom izvješću, EBRDova povlaštena dionica AKGR-P-A je pretvorena u običnu dionicu, što bi moglo značiti da je EBRD izašao iz Agrokora, iako njihovo izvješće kaže da imaju još 2.31% (EBRD je originalno imao nešto preko 8%). Prema Agrokorovom izvješću za 2015., Ivica Todorić ima 95.52% kompanija založeno u Nizozemskoj firmi preko koje je podignuo PIK kredit. To je samo jedan postotni poen iznad 94% iz PIK ugovora što znači da je vjerojatno puno manje od predviđenih 120M EUR utrošeno na otkup EBRDovih dionica.
• Ostatak novca od PIK Kredita (preko 200 milijuna EUR) izgleda da nije ušao u Agrokor niti je iskorišten za kupnju Agrokorovih dionica već je ostao u vlasništvu Ivice Todorića.

Jedna pikanterija izranja iz Agrokorovih godišnjih izvješća 2009–2013: životni stil najbogatije hrvatske obitelji košta oko 60 milijuna kuna godišnje. To je iznos dividende koja je isplaćivana u tom periodu. Ova teza je dodatno podržana jednim od uvjeta PIK Toggle kredita: maksimalni dozvoljeni iznos isplate dividendi iz Agrokora tijekom trajanja kredita je 8 milijuna eura godišnje.

Nakon preuzimanja Mercatora, iznos isplaćenih dividendi naglo raste, na 280M kuna u 2014 i 190M kuna u 2015. Agrokor je efektivno počeo servisirati Todorićev kredit.

Zaključak

Agrokor nije propao radi preuzimanja Mercatora već zbog nagomilanih dugova koji nisu bili otplaćivani već samo odgađani po skupim kamatama. Suočen s ograničenim tržištem i jakom međunarodnom konkurencijom od koje je polako ali sigurno gubio tržišnu utakmicu, Agrokor je pokušao rasti akvizicijom konkurencije, no kombinirani dugovi Mercatora i Agrokora su samo ubrzali neminovan proces propadanja Agrokora.