The one about Andrea Latino
A free-for-all
Friends,
I hope that all is well with you and yours, and that this e-mail finds you on a boat with shoddy connection, in the tropics, three months after I sent it.
Now accepting keynotes for 25Q3-25Q4
The 2025 lineup:
What to do when you don’t know what to do. How to create a sustainable competitive advantage through superior adaptability. (Based on the new book by the same name.)
Succeed big, fail small. How to drive efficiency and effectiveness of innovation at scale.
Managing uncertainty. How to steer an organization through a sea of change.
Retail 3.0. How to build a profitable retail business in the modern marketplace. (Based on the 2025 follow-up to the 2022 smash hit white paper The Gravity of e-Commerce.)
If you want to book me for your event, corporate speaking slot, or workshop, merely send me an email. To make sure I am available, please do so at your earliest convenience; my availability is limited and the schedule tends to fill up fast. More information may be found here.
The TL;DR
Experimentation is not a department found somewhere between “R&D” and “Business” on an organizational chart, but rather a lens through which every decision should be seen and measured.
Every test is a snapshot of a unique context.
Firms looking to drive innovation should create a culture in which the question is not “should we test this?”, but “why have we not tested this yet?”.
Snoring can cure appendicitis. Potentially.
Personal updates before we go-go
As some of you may have observed, I recently wrote on social media that this newsletter was in danger of not going out. The reason, quite simply, is that this spring keeps shitting on me from great heights. I was in ER, waiting for surgery.
On Monday, I started to get stomach pains. Thought no more of it; assumed I had just eating something I should not have. On Tuesday, the pains got worse. Come afternoon, they went from bad to searing. A few hours later, they all focused on one particular part of the abdomen: the appendix.
Off to the ER we went. Sure enough, the doctor drew the same conclusion. Appendicitis.
Alas, the local hospital did not have the requisite personnel nor equipment, which meant I had to go to another (larger) hospital and spend the next 14 hours undergoing even more tests, scans, and whatnot. Ultimately, the decision was made that I should have surgery to have the appendix removed.
However, since the surgeons were busy, I had to be put in a room. And because of the aforementioned spring’s proclivity towards droppings, I ended up in one not in the emergency ward but in the otorhinolaryngology ward. For those unaware, that is the ward that specializes in the ear, nose, throat, base of the skull, head, and neck. Which includes people who suffer from sleep apnea.
So, yes, I found myself in a room with a - I kid you not - two meters tall, 160+ kg heavy man who snored so loudly that the paintings fell off the walls. I have never heard anything like it in my life and pray I shall never again.
Anyway, apparently the bacteria could not take the thunder rolling through the building either, so the pain eventually subsided. By next morning, it had gone down so much that the surgeons reconsidered. Eventually, they decided to send me home.
So now, at long last, I am back in my house. There is still some pain, but very little compared to what was.
Right now, it is a waiting game. If the inflammation goes away completely, great, though I should be aware it is will most likely be temporary (whatever time period that translates to). If it increases again, well, straight to surgery I will go. Heads up.
A massive thank you goes to all the people who sent me their well wishes. It was, is, and forever will be much appreciated.
As a result of my time in hospital, today’s newsletter will not feature the usual market analyses; I simply have not had the time nor energy to do any (I only opened my laptop for the first time this week two hours ago). I can only apologize. Normal services will resume next week.
Moving on.
The expert interview: Andrea Latino
Strategic experimentation, part IV
Andrea and I first met a few months ago at the Richmond Digital Marketing & AI Forum (at The Grove Hotel just outside of London). We were both there to speak and, by sheer coincidence, ended up next to one another at the breakfast table. It soon became apparent that we not only shared a view on a great many things to do with strategy and innovation, but also a sense of humor and a strong view of what is and is not suitable to eat first thing in the morning.
Andrea is, without a shred of hyperbole, one of the most impressive people in innovation that I know. And so, it is with great pleasure that I am able to feature him in this week’s interview.
For those readers who, for some reason, do not yet know you, could you tell us a bit about not only about who you are, but also how you ended up where you are?
I suspect most readers can be forgiven for not having my name on speed-dial.
I’m currently working as Global Digital Lead at Nestlé, the world’s largest FMCG company. The team I am part of - Innovation Acceleration - sits in R&D, but benefits from a strong business connection.
We run an Accelerator program that takes embryonic product ideas and, in just six months, brings them to a controlled shop test. That speed forces us to rely on a rigorous “digital experimentation” methodology to de-risk value propositions before our colleagues in the business decide to fully launch them (or not). The approach has become a contributing force to cross-category pipelines.
Prior to Nestlé and to my last 5+ years in Switzerland, I enjoyed a decade as a solopreneur consultant and entrepreneur in my home country of Italy, helping C-suites across industries with their “biggest digital problem” (which usually turned out to be three problems, plus culture). I also survived Italian public service, and had brief stints in academia, journalism and executive training.
After advising corporations for years I felt a growing itch: what would it be like to own the problem end-to-end instead of stopping at the consulting level? Nestlé offered me that chance and the opportunity to see my impact amplified at scale. It’s not easy, but I still haven’t got used to that feeling of impact at scale when it happens.
It is safe to say that you are an experimentation expert. Most companies think of experimentation as a part of an R&D process in which the end goal is to launch a market-ready product. What is your take?
I like to say that experimentation is not a department tucked between “R&D” and “Business” on an org chart; it is a lens through which every decision should be seen and measured. Whenever money, reputation or scarce time is on the line, the rational first move is to ask, “what small cheap test could save us from a catastrophic large, expensive mistake?”.
Media budgets, especially digital ones, are my favorite cautionary tale. Give a traditional trade marketer recently promoted to management also “that digital stuff", and half a million euros, and the easiest thing that will happen is that he or she will buy every shiny channel suggested by an agency PowerPoint, only to later pray to the gods of CPM that the ROAS or MER will not be THAT negative (assuming he or she has bothered to study what those metrics are).
A smarter approach for our newly promoted friend is to slice that agency budget for controlled experiments in iteration waves that pit message against message, audience against audience. Within days, reality rudely punctures assumptions, and he/she can double down on what genuinely works instead of what felt persuasive in a brainstorm. Then he/she can scale the spending and repeat the process as our campaigns lose steam.
This is nothing new or rocket science in the performance digital marketing or in eCommerce - I would say it’s the basis of operating in the filed - but I’m still baffled by how little this has spilled into other parts of larger organizations.
The same logic applies to physical products. Veteran insight managers inevitably possess intuition, yet betting solely on gut feeling and Mintel reports against the messy complexity of consumer desire is like bringing a knife to a gun fight. Yet if you’re willing to evolve, marrying classic declarative CI with behavioral digital experiments - think split testing landing pages or limited-batch drops - companies let actual buyers vote with email addresses and/or wallets rather than with paid survey answers.
This mindset, however, does demand fresh skills, data literacy, and occasionally a bruised ego when beloved ideas flop in the wild, but the alternative is far costlier: slowly incubating a potential failure.
Statistically speaking, most employees (whether long-toothed veterans, newbies, managers or shop workers) have no idea whether an experiment will work or not. But should you always test before you launch?
The honest answer is that the desirability of testing rises and falls with your appetite for risk. If the company is standing on a fiscal cliff - think a make-or-break turnaround year - then treating every cent like an endangered species is simply prudent stewardship, and testing becomes non-negotiable.
Conversely, if you enjoy Apple-level margins or face a now-or-never market window, you might swallow a higher probability of failure because being first matters more than being certain. Yet even in those scenarios, embedding at least a pocket of experimentation - perhaps on pricing, messaging or distribution - tends to pay for itself.
My rule of thumb is that structured learning should be the default and skipping it should require "written justification", the way surgeons must account for any left-behind sponges. When testing is embedded as the cultural norm, it sheds its reputation as a hurdle on a checklist and becomes like the air people breathe.
A lot of what is written about experimentation and testing comes from software, but conventional wisdom says that you cannot use the same principles in physical NPD. I know your view is slightly different. Could you elaborate?
Conventional wisdom in this case is several releases behind the current operating system, if you allow me the software metaphor. Digital tools now allow physical innovators to mimic the rapid learning cycles once exclusive to software releases.
Take Kickstarter as an example: prospective consumers routinely pledge cash for products that exist only as a slick render and a charismatic founder video. However, if enough consumers bite, the creators earn the funds to proceed; if the campaign fizzles out, nobody bankrupts a factory. That is market validation dressed up as crowdfunding. At Nestlé we borrow some of same principles, minus the pledging part.
Where do most people go wrong with experimentation?
The gravest sin is mistaking experimentation for clairvoyance.
A/B tests will not predict the next decade of consumer trends any more than a weather app will tell you which day to retire. Their purpose is narrower: to test assumptions, collect evidence and help you make decisions. Trouble begins when teams treat a positive result as pure gospel, forgetting that statistically significant (assuming you can even get that) does not equal commercially significant, as the only true test of Product-Market Fit for an unreleased innovation is a full scale launch.
Every test is a snapshot of a unique context: product, price, creative, season, even the phase of the moon if you market fishing gear or biodynamic foods (if you always believe intuitively that "it’s even better than organic, right?", you are sorely mistaken - it’s esoteric stuff with zero scientific backing).
Companies keen on building cross-category benchmarks must restrict experimentation to tightly controlled templates to reduce potential variability, but that defies the flexibility spirit of the whole approach. Finally, process discipline - documenting hypotheses, sample sizes, decision rules - is often skipped in the rush for a quick answer, turning what should be a process inspired by science into anecdotical evidence at best with a few slides attached.
What is the single biggest practical tip you would tell organizations looking to experiment?
Treat experimentation as infrastructure, not as an occasional stunt. That means establishing governance so people know who can green-light tests and who owns acting on the results, hiring or upskilling T-shaped digital specialists; investing in an analytics stack that reports insights in human language rather than hieroglyphics, and - perhaps most overlooked - agreeing upfront on what success looks like and how bold you are willing to be if the data whisper “pivot, you fools!”.
Once that scaffolding is in place, begin super small: one landing page, one SKU, one experiment. The first practice will earn you experience and political capital for bigger bets, and soon the question in meetings will hopefully shift from “should we test this?” to “why haven’t we tested this yet?”. And if all of that sounds about as daunting as rewiring an airplane mid-flight, my shameless plug is here: there a JP guy I know who could probably help you jump start this. Maybe give him a call.
And with that, we thank Andrea not only for taking time out of a hectic schedule to answer our questions, but also the very generous plug which I will have to repay with a cocktail in the near future.
Next week, Doug Garnett and I will continue our deep dive into experimentation as a way of navigating uncertainty. Until then, have the loveliest of weekends.
Onwards and ever upwards,
JP