Thinks 2009

Marina Nitze:”Crisis engineering is harnessing a crisis to make rapid transformational change. A complex system is any system that’s made of humans and computers. The computer part is generally pretty easy to change. And the human part is very difficult to change. But if certain crisis conditions are present, you can transform the human part of a system very rapidly. Crisis engineering is about how to recognize those indicators and then harness that moment to make rapid change against the human part of your system, and probably the computer part, too.”

FT: “Emboldened by big leaps in AI’s software programming capabilities over the past six months, tech industry leaders and researchers are increasingly confident that an AI that can improve itself with little to no human input is within their grasp. These self-taught AIs could keep on building new, more powerful versions of themselves. Successive generations could repeat the trick over and over again, adding new capabilities or making themselves run more efficiently. This flywheel effect, known in the industry as “recursive self-improvement” (RSI), could quickly lead AI beyond the large language models that underpin Google’s Gemini, OpenAI’s ChatGPT and Anthropic’s Claude, into uncharted territory. AI’s optimists believe it is the key to so-called superintelligence, or the point at which AI surpasses the abilities of the human mind.”

David Oks: “Why did China get rich, and India didn’t? What explains the Sino-Indian divergence?…China’s explosive growth wasn’t simply a matter of “freeing the markets,” reducing the role of the state, and announcing that it was now glorious to get rich; nor was it simply a matter of government intervention to support the manufacturing sector and subsidies for favored companies. China succeeded because it spent decades on the basics of human development and social modernization. India did not. The rest is just commentary.”

Ben Thompson: “How many companies could actually employ that cash in a way that generated a high rate of return? It’s hard to imagine a better option than Google. The company is not only investing in AI, but has optionality in terms of outcomes: its Services business benefits from the investment, it is in contention at the model layer with Gemini, and it can sell capacity to the frontier labs. Moreover, that capacity has a sustainable cost advantage because of TPUs, which means that in a world where compute becomes a commodity — as hard as that is to imagine right now — Google is the hyperscaler that is poised to make the most profit.”

The Intent You Rent Back (Part 6)

The three states of post-purchase intent.

There is no single solution to post-purchase intent because there is no single intent state. After a purchase, a known customer sits in one of three states, and each needs a different owner. Mistaking which state a customer is in is how brands reach for the wrong tool and end up at the auction.

Engaged — manufacture owned intent

If the customer still opens, clicks, taps, browses or replies, the brand does not need to detect intent from outside. It needs to create a surface where intent can keep appearing inside. That is the job of owned attention. A conventional promotional email waits until the brand wants to sell; a NeoMail earns attention before the brand needs to sell, giving the customer a reason to open — a useful idea, a quiz, a recommendation, a small reward, a moment of recognition. Over time the inbox becomes a living surface rather than a dumping ground for offers. And every meaningful action — a click, a tap, a saved item, a preference, a reply — becomes telemetry: identified, first-party signal about what the customer is thinking and may need next. Owned attention is the only owned intent signal, which is why attention sits upstream of transactions. This is Atrium’s first job: manufacture low-cost, continuous, identified attention so that intent never has to be rented later.

Predictable — anticipate before adtech detects

Some intent does not need detecting at all, because it can be anticipated. Replenishment cycles, renewals, seasonality, usage cadence, life events, festivals, travel, expiry dates and household rhythms all create predictable windows. A brand may not know the exact day a customer will search, but it can know the likely week or trigger. This is where intelligence earns its place. A BrandTwin, fed by the Customer, Product and Decision-Trace Context Graphs and the relevant world data, models the next likely intent window and acts before the customer ever appears on Google. The shift is from catching the signal to anticipating the window. A pet-food brand should not wait for dog food delivery to be typed into a search bar; an insurance brand should not wait for the renewal comparison to begin. If the timing is modellable, the brand should move first. This is Meridian’s job: convert predictable re-intent into lifetime value before that intent becomes expensive.

Silent — route through the network

The hardest case matters most. The customer is known but silent: no opens, no clicks, no visits, no replies. Owned attention has failed and prediction is too coarse — yet the customer is in-market somewhere else. This is the moment adtech was built to monetise. But there is another possibility: your silent customer is alive on someone else’s surface. They are reading a partner brand’s email, using a partner app, engaging with a non-competing brand in the same life moment. Their attention is not gone; it is simply not with you. NeoNet turns that fact into a recovery system. The old answer was an auction — upload the audience, bid for the user, pay the tax, hope the platform finds them. The NeoNet answer is routing — recognise that a known-but-silent customer has re-engaged on a trusted partner surface, and route a relevant recovery message through that surface before the brand falls through to paid retargeting. The network need not share raw data; it shares a decision — this person, silent for Brand A, is active in a trusted context where a relevant Brand A message can be shown. Recovery stops being a retargeting problem and becomes a routing problem. And routing should always be cheaper than bidding.

Thinks 2008

NYTimes: “Scientists at Columbia University have edited the DNA of early human embryos with unprecedented accuracy, an achievement that could open the way to babies engineered with particular characteristics. The prospect has fueled controversy for years. On the one hand, the technology might one day enable parents to safely repair disease-causing mutations in embryos. But it might also be used to select desired traits — a practice that some ethicists have argued is nothing short of eugenics. Dieter Egli, a geneticist at Columbia University who led the research, called for a public conversation about the pros and cons of altering embryonic DNA. “As a scientist, you can provide the data for discussion, but then essentially there you stop and let others take over,” he said.”

FT: “For most people working in markets, business or government, the canonical examples of a global financial crisis are 1929 and 2007-09. Liaquat Ahamed’s new book brings to life a third global cataclysm that was not like either of its successors and raises intriguing parallels with today’s technology and geopolitically charged world. The cast of 1873 includes Mark Twain, Karl Marx, Egyptian accountants and Prussian officers turned speculators. The boom and the bust they lived through helped create the modern world.”

WSJ: “Finance chiefs are trying to get a better read on how much AI their companies are using to avoid a sticker shock moment as vendors begin charging for the technology by tokens. The shift to pricing based on usage, and measured by tokens—the basic unit of measurement for AI computing—is creating new challenges for even the most experienced finance teams. CFOs used to paying flat amounts for technology are finding costs more unpredictable and harder to model as they build agents and embark on ambitious AI investments. Twenty-six percent of companies say they have a comprehensive view of their AI costs, while 50% have some visibility and 22% report no visibility or visibility after billing, according to an as-yet-unreleased survey from KPMG. “It’s a new resource that needs to be managed that didn’t exist quite that way, and we’re seeing exponential growth,” said Steve Chase, KPMG’s global head of AI.”

Mint: “India already sees $15 billion in annual philanthropic giving, nearly 100 times the annual need estimate. This is comparable to India’s startup funding, placed at around $12 billion dollars every year. Yet, 80% of non-profits struggle to scale due to funding constraints. The issue is not capital but a lack of ‘flexible funding.’ Flexible funding, akin to venture capital for startups, is a distant dream in the non-profit sector. Most funders restrict grants to specific programmes, tightly defined budgets and short timelines. We have argued previously that building scalable, low-cost solutions requires iteration, experimentation and adaptability—conditions incompatible with line-item funding.”

The Intent You Rent Back (Part 5)

A worse signal than the one you should own.

There is a final insult buried in the economics. The intent a brand re-buys from adtech is not just expensive; it is a worse signal than the one the brand could have owned.

The adtech signal is probabilistic, delayed and anonymous. The platform infers that someone may be in-market by matching devices, cookies, cohorts and behaviours. It does not tell you what it knows. It rarely hands over the underlying customer truth. It sells you an audience and then charges you, again and again, to reach it. You are buying a guess about a stranger who is probably your customer — an audience you must keep renting, because you never take possession of it.

The signal a brand could have owned is the opposite on every axis: deterministic and identified. This is your customer. This is their email and mobile. This is their order history and their value. This is the last meaningful attention event, and this is the likely next window. This is the best route to the next transaction. Nothing is inferred, because nothing needs to be — the relationship already exists.

Paying a premium for a downgrade

Put the two side by side and the absurdity is plain. A brand pays a premium for a poorer version of a signal it had the right to build for nothing. The re-bought customer arrives stripped of identity, with no history attached, mediated by a platform that keeps the customer truth for itself. The brand has effectively handed its own customer knowledge to the auction and is now buying a thin slice of it back at a markup. That is the tragedy of AdWaste, stated at the level of the signal: not merely that you pay twice, but that the second thing you buy is worse than the first thing you gave away. Worse still, the brand loses the learning: every recovery handled by the auction teaches the platform about your customer and teaches you nothing.

The problem is sequencing, not adtech

None of this makes adtech the villain. Adtech is genuinely useful for what no owned or shared system can see — a true new prospect, a customer active only on surfaces the brand cannot reach. The problem is not that adtech exists; it is that brands use it as the first detector of repeat intent instead of the last resort. NeoMarketing simply reverses the order. Spend adtech last. First own the intent you can see, because an engaged customer is generating it on your surfaces right now. Then predict the intent you should expect, because much re-entry is modellable. Then share the intent your partners can see, because your silent customer is often alive on someone else’s surface. Only after those three fail should you rent intent from the auction — and then for the narrow slice that genuinely nothing cheaper could reach. The next three parts take those three owners in turn, beginning with the engaged customer. Reversing the order is not a tactic; it is a change in what marketing treats as its first instinct.

Thinks 2007

Aakrit Vaish (Activate): “AI does not kill IT services. It collapses the delivery pyramid, splits pricing along the bespoke/commodity line, and converges the whole industry on professional-services economics. The margin trapped in headcount-heavy managed services and implementation is being released. The firms that capture it will not be better staffing businesses, but vertically deep, partner-led, agentic delivery firms that own workflow completion rather than workflow effort, operating as the intelligence layer inside enterprise GCCs. The $98.4B GCC ecosystem is the distribution. The agent stack is the engine. Run-time is the real estate. And India, with the domain depth, the talent cost structure, the enterprise trust, and the GCC density already in place, is holding the deed.”

NYTimes: “Consultants and executive coaches who don’t have the bandwidth to address every inquiry are referring some clients to their A.I. doubles. Harvard Business School professors have incorporated A.I. versions of themselves into courses and office hours. And executives are using their A.I. avatars to address employees in other countries in their own languages. Whipping up an A.I. chatbot or avatar is easy. Allaire built his using Claude. A handful of start-ups provide interfaces that make it even easier and offer more control: Delphi takes your content and instructions and creates a voice and text chatbot that mimics you, while A.I. video generators like HeyGen and Synthesia will do the same for a digital avatar that copies your appearance.”

WSJ: “Run-A-Muck’s bigger plan is to see which stories land with readers and turn them into the backbone of other money-making projects. Short stories from Hemingway’s tales to “Brokeback Mountain” and 2017’s “Cat Person” have long been source material for movies. Run-A-Muck thinks they could also expand into full-length novels, podcasts, TV shows, immersive events, digital shorts, microdramas and other vertical-video formats. It also hopes to flip the script, so to speak, publishing short stories based on new and upcoming series and films…The company is betting that the generation raised on a diet of YouTube channels and Instagram captions will also embrace the novel’s shorter sibling.”

FT: “Under-30s make up about half of India’s 1.4bn people — the world’s largest youth population — and according to a March report by Azim Premji University in Bengaluru, nearly 40 per cent of graduates aged 15-25 and 20 per cent of those aged 25-29 are jobless. While those unemployment rates are higher than for those less educated, many young people and their families see success in public exams as their best hope for economic advancement and security. Every year about 200,000 late-teen students travel by train, bus, car or motorbike from all over India to cram in Kota, a city in Rajasthan.”

The Intent You Rent Back (Part 4)

The purchase starts the silence.

A purchase feels like the end of a successful journey. In reality it is the start of the next-purchase clock — and brands consistently misread the moment.

The customer got what they came for. The immediate need is satisfied. They stop browsing, stop opening, stop visiting. They may not need the brand for weeks or months, and the better the purchase experience, the more complete the closure. So the brand goes blindest at exactly the moment it should be watching most carefully. Many CRM programmes enter a quiet zone after purchase: an order update, perhaps a review request, perhaps a replenishment nudge, then silence. If the customer keeps engaging they stay visible; if they stop, the brand quietly files them as inactive.

Relationship silence is not market silence

This is the costliest confusion in marketing: brands mistake relationship silence for market silence. Inactivity in your channels does not mean inactivity in the market. It only means the customer is no longer paying attention to you. A customer who does not open your email may be browsing a competitor. A customer who has not opened your app may be searching the category. A customer who ignores your WhatsApp may be comparing options on a marketplace. To martech, all three look dormant. To adtech, all three are warming up. The customer has not gone cold; they have gone elsewhere — and elsewhere is precisely where your owned systems cannot follow.

How an owned customer becomes a rented one

That confusion is the mechanism by which an owned customer becomes a rented one, and the sequence is mundane and expensive. The customer buys. Engagement falls, as it always does after a purchase. The brand reads the fall as disinterest and eases off — or worse, keeps up the same promotional pressure and trains the customer to ignore it. The real re-entry, when it comes, surfaces off-property, where adtech sees it first. The platform packages it as a retargeting audience and sells it back. The brand pays the tax and books the result as performance, never noticing it just re-bought a customer it already owned and had merely stopped watching.

Watch attention, not transactions

The fix begins with measuring the right thing. The next-purchase clock should not run on the order book; it should run on attention. A brand that watches days-since-last-transaction will always be late, because by the time the transaction gap is visible the attention gap has already done its damage. A brand that watches days-since-last-meaningful-attention has an early-warning system — it can see the drift from engaged to weakening to silent before the revenue stops, which is the only window in which intervention is still cheap. The post-purchase period is not a lull to be left alone. It is the most important surveillance window a brand has, and the one most programmes sleep through. Treat the quiet as the signal, not the absence of one. Most dashboards are built to report the transaction gap; almost none are built to report the attention gap, which is why the damage is usually found too late to be cheap to fix.

Thinks 2006

David Oks: “Let’s say you run a factory. You decide that you want your lines to produce fewer defective goods: maybe you want to improve your yield from 95 percent to 98 percent. So you decide to invest in better training for your workers: maybe training now lasts six weeks instead of two weeks. This works, and now your yield is higher; but that change makes other things more attractive too. For example: now that your yield is higher, it makes sense for you to reduce your inventory, since fewer defects mean you no longer need a large buffer of spare parts to replace the bad ones. So now you’ve cut your inventory: but now it makes sense for you to shorten your production runs and switch more frequently between products, since without a mountain of inventory to work through you can afford to change what the line is making. And if you’re switching frequently between products, then it makes sense for you to invest in flexible, reprogrammable machinery instead of dedicated, single-purpose equipment. So one relatively small tweak shifts the entire calculus of what you do. In short: each practice makes the others more valuable, and each practice is valuable because it’s implemented alongside other complementary practices. Doing just one of these things—investing in flexible machinery, for example—doesn’t really make sense alone. The practice needs to work well with all the other practices that you have.”

Eric Ries on the key takeaway from his new book Incorruptible: “That it’s entirely possible to build a company that is both profitable and also stays true to its purpose. We’ve all unconsciously absorbed an orthodoxy that says that these attributes are opposed. The truth is that the most valuable companies are the ones that build trust with everyone they touch, customers, employees, and investors alike.”

WSJ: “Silicon Valley venture-capital firms are desperate for bets that can survive—and thrive in—the AI reckoning. Investors known for early investments in software, internet services and social-media companies like Snap and Uber have begun venturing far outside of their comfort zones into investments in physical technologies and materials tied to the artificial-intelligence boom. They are making new wagers on AI infrastructure like chips, power and manufacturing, as well as a far-ranging category called physical AI, or autonomous machines that can understand and perform complex real-world tasks. Venture-capital investment in global robotics and physical AI grew to $26 billion in 2025 from $4.2 billion in 2019, according to PitchBook data. This year, companies in those sectors have already raised more than $23 billion as of May 20.”

FT: “China’s ability to mass manufacture a wide range of goods at low cost stems from the governance that enables it. Centralised control allows Beijing to easily mobilise resources and co-ordinate measures — across provincial governments, state-owned banks and regulatory bodies — to meet its policy goals. Limited democratic accountability also gives the Chinese Communist Party the ability to pursue long-term industrial policy and sidestep opposition, such as planning disputes. China combines this central power with intense decentralised competition. Regional officials and enterprises compete for state backing, which is often conditional on performance, creating strong incentives to prioritise output and innovation ahead of profits. This model of state-led capitalism has been refined over decades and has underpinned the country’s ability to nurture entire industries, scale and develop dense, vertically integrated supply chains.”

The Intent You Rent Back (Part 3)

Martech is blind to the wrong intent.

It would be too easy, and not quite true, to say that martech is blind to intent. Martech sees a great deal of intent — just not the kind that matters most, at the moment it matters most. The distinction is worth getting exactly right, because it determines where a brand looks for its next customer.

Martech sees declared intent: preferences, surveys, wishlists, sizes, categories, loyalty choices. It sees historical intent: what the customer bought, how much they spent, when they last transacted, what their basket contained. And it sees on-surface intent: what they clicked, browsed, searched, added to cart or abandoned on the brand’s own property. This is valuable — it is the foundation for personalisation, replenishment, recommendations and lifecycle marketing. For an engaged customer it is a rich picture.

But it is not enough, because the most valuable intent a brand can act on is rarely the intent expressed inside the brand’s own walls. It is the re-entry signal: the moment a past customer begins thinking about the category again after a period of silence. A skincare buyer starts researching sunscreens before summer. A grocery customer’s household needs change. An insurance customer begins comparing renewals. A jewellery customer starts browsing gifting ideas before a festival.

Where re-entry first appears

Where does that intent first appear? Almost never in the brand’s own email, app or website. It appears in search, in social, in marketplace browsing, in comparison content and review sites, in creators and partner brands, in payments and travel and adjacent categories. It leaks into the world long before it shows up on a surface the brand controls. That is the gap. Martech sees the intent a customer chooses to express on a surface you own; adtech sees the intent that leaks into the world — and the most expensive version of that leakage is when the customer is already known to you. The richer a brand’s owned data, the easier it becomes to forget how partial it is.

The gap, stated precisely

So the problem is not that martech is blind to intent. It sees plenty of it: what a customer has told you, what they have bought before, and what they do on your own email, app and website. What it cannot see is the intent that shows up somewhere else — a past customer starting to look at the category again on a marketplace, in a search, or on a rival’s email. That blind spot is not about working harder or buying better tools. It comes down to where the data is created: a CRM can only record what happens on your own surfaces, so it cannot watch a customer comparing prices on Amazon or reading a competitor’s newsletter. Better segmentation will not fix it, because the gap is one of coverage, not analysis.

This is exactly why the two costliest customers are the ones martech loses sight of: the one who has just bought, and the one who has gone quiet. Both, almost by definition, do their next-purchase thinking somewhere you cannot see. A brand that mistakes its rich view of on-surface behaviour for the whole picture will keep being caught out — most painfully by its own former customers, walking back into the market without it noticing.

Thinks 2005

Business Standard: “India achieving developed-country (Viksit Bharat) status by 2047 would require appropriate structure and strategy besides getting its act together on both types of investments — physical capital and human capital. India Out of Work diagnoses the key adverse trends threatening to derail the country’s much-acclaimed demographic dividend. India, according to the authors, is beset with a poly-crisis on triple fronts — employment, education and economy.  India needs to grow at 8 per cent consistently to reach advanced-country status by 2047. It has to swim against the tide of growth, which has fallen from 7.8 per cent during 2004-14 to an arguable 6.2 per cent between 2014 and 2024. Achieving this status would require an annual net job creation of 10-12 million in the non-farm sector, which has fallen from 7.7 million to 4.3 million.  The other handicaps are India’s high population density of 483 per sq km, against China’s 148 in 2023; a low employed-to-population ratio of 41 per cent (against a global average of 56 per cent); and dominance of the less-efficient informal sector, which accounts for 85 per cent of output and employment, compared to 60 per cent globally. “

Frank Bruni on brain health: “Along with popular brain journalism there are plentiful brain books, written by members of a growing pantheon of brain whisperers who promise that the right diet, exercise and engagement can safeguard our smarts. In “Keep Sharp,” Dr. Sanjay Gupta assembles tools for a task detailed in the subtitle: “Build a Better Brain at Any Age.” “The Ageless Brain” presents a best-brain protocol by Dr. Dale E. Bredesen, and it inspired a recipe collection, “The Ageless Brain Cookbook for Seniors,” by Hadwin Macy.  Dr. David Perlmutter has stretched his prescriptions for brain health into more than a half dozen volumes, including “Grain Brain” (about the danger of too many carbohydrates), “Brain Maker” (about the benefits of gut microbes), “Brain Defenders” (about the importance of the immune system) and “Brain Wash” (about detoxing the brain). To feed your thoughts, heed his thoughts. Or just play with puzzles.”

WSJ: “AI is coming to small business, helping companies to organize supply chains, plan production and execute other functions in ways that only multibillion-dollar enterprises were once able to afford. More than 50% of small businesses, including owner-only firms, said they plan to use AI tools this quarter to boost productivity, according to a March survey from Citizens Financial. Their use of AI is rapidly increasing. Among businesses with 20 to 99 people, 84% said they plan to use artificial intelligence this quarter. And 91% of those with more than 100 workers said they are using it to boost productivity right now.”

FT on spiralling: “At the risk of sounding platitudinous: life can be horribly frustrating. You feel like you’re finally getting somewhere, and then you mess things up again (in the way you always do) and you’re right back at square one, stuck in your familiar, depressing rut. Round and round you go in the same old circle, never seeming to make any progress, until you finally lose that most precious of commodities: hope. But consider this: what if you were going round and round, but in a way that is actually OK — normal, healthy, desirable, even? What if, despite all your cock-ups and self-chastisements, you were actually getting somewhere, just on a slightly more roundabout trajectory to the one you had imagined?”

The Intent You Rent Back (Part 2)

Presence, not intelligence.

The common explanation for adtech’s dominance is that it is simply smarter than martech — better algorithms, better bidding, better optimisation, better models. That is partly true and entirely beside the point. Adtech does not win because it understands the customer better. It wins because it is present at the moment intent appears.

Search knows when a customer starts looking again. Social knows when behaviour changes. Marketplaces know when comparison begins. The open web knows when browsing resumes. Retail media knows when category interest returns. Adtech sits across the surfaces where people live their commercial lives, so it feels the tremor before the brand feels the earthquake.

Martech, by contrast, sits almost entirely on surfaces the brand owns: email, app, website, WhatsApp, loyalty, CRM. It knows what the customer has done with the brand. It rarely knows what the customer is doing everywhere else. So a customer can go silent in your CRM and reappear, weeks later, as a paid retargeting audience you bid to win back. The customer did not vanish. They moved their attention elsewhere. Martech stopped seeing them; adtech never did.

The product is the answer, not the ad

This is why adtech’s real product is not the ad. The ad is the delivery mechanism. The product is the answer to one question: who is in-market right now? A platform that can answer that across billions of sessions has built something genuinely valuable — and genuinely rentable. It can charge you, again and again, to be told a fact about your own customer. That is a structural advantage, and structural advantages are beaten by changing the structure, not by bidding harder inside it.

Why the distinction decides the answer

The difference between presence and intelligence is not academic; it decides the entire shape of the solution. If adtech won on intelligence, the response would be to build a cleverer model and out-think it — and you would lose, because the platforms have more data and more compute. But because adtech wins on presence, the response is different and winnable: be present where intent surfaces. Be present on your own property, where an engaged customer still shows up. Be present across a cooperative network, where a silent customer is alive on a partner’s surface. The fix is surface and cooperation, not a smarter algorithm. That reframing is liberating for a CMO. You are not in an arms race against Google’s models; you are in a coverage contest about where you can see your own customers — and most of that coverage you can build or share rather than rent. The platforms do not read intent better than you could. They are simply everywhere the customer goes. Match the coverage and the monopoly thins — the very gap this series exists to close. It also explains why the usual fixes fail: a better DSP, a sharper agency or a bigger budget all optimise the renting; none of them closes the coverage gap that forced the rent in the first place.