De Beers just paused production at Venetia, its flagship South African mine, for two years. Rough diamond prices are down roughly 50% from their 2022 peak. A company that has spent a century controlling global diamond supply and invented “a diamond is forever” got blindsided by a shift in consumer intent it never saw coming on its balance sheet until it was too late to react.
This is not a luxury goods story. It’s a data story, and it should terrify anyone who runs a brand on quarterly revenue dashboards.
De Beers had excellent financial reporting. Sales figures, margin data, sightholder allocations, all tracked meticulously, all lagging by months. What it didn’t have was a system built to catch consumer sentiment moving in real time, long before that sentiment showed up as a dent in revenue. Lab-grown diamonds didn’t ambush the market overnight. They eroded it steadily: in search behaviour, in social conversation, in shifting definitions of what “value” even means to a 28-year-old buying an engagement ring. By the time it hit the income statement, the disruption was already structural.
For CMOs and growth leaders, this is the case study you didn’t ask for but need anyway. If your marketing analytics early warning system is built entirely on lagging indicators, you will find out about your “De Beers” moment the same way De Beers did: after it’s already cost you the market.
The details of the diamond slump are worth sitting with, because they map onto categories far removed from jewellery. Weak consumer demand in China, a generational shift toward gold and other investment assets over discretionary luxury, and lab-grown supply maturing faster than anyone in the natural diamond trade modelled, all of it compounded quietly for roughly two years before De Beers cut official rough prices and idled a mine that cost $2.3 billion to build. Anglo American has written down De Beers’ value multiple times, actively divesting its 85% controlling stake in De Beers. None of that happened overnight, and none of it was invisible in the data, just invisible in the revenue data.
That’s the pattern every marketing leader needs to internalise. Disruption rarely arrives as a shock. It arrives as a slow leak that only becomes a crisis once it’s large enough to show up where finance is looking.
The Lagging Indicator Trap
Let’s define terms, because the distinction matters more than it sounds like it should.
Lagging indicators are outcomes. Revenue, quarterly sales, market share reports, churn rate. They tell you what already happened. They are accurate, auditable, and board-ready, and, by definition, backwards-looking.
Leading indicators are behaviours. Search intent, social sentiment velocity, category consideration rates, price sensitivity signals. They tell you what’s about to happen. They are noisier, harder to standardise, and much less comfortable to present in a QBR, which is exactly why so many organisations underinvest in them.
Why high-margin categories are the most exposed
Here is the part that should worry premium and luxury brands specifically: the higher your margin, the more room a disruptor has to undercut you on price while still building a profitable business. Diamonds carried enormous margin built on manufactured scarcity and emotional narrative. Lab-grown alternatives didn’t need to match that margin; they just needed to be “good enough” at a fraction of the cost.
The same math applies to premium SaaS, designer goods, high-end financial products, and any category where price partly reflects brand mythology rather than functional differentiation. Lagging vs leading indicators debates in marketing are not academic in these categories; they are existential.
The micro-shift problem
Revenue doesn’t move in a straight line down. It holds steady, then steady, then it cracks. Because consumer definitions of value shift gradually and unnoticeably before they shift the profit and loss suddenly and visibly.
By the time De Beers’ sales numbers reflected the lab-grown shift, it had been building for years in:
- Search behaviour comparing “natural” vs “lab grown” diamonds
- Social sentiment around ethical sourcing and traceability
- Price-per-carat comparison content proliferating across TikTok and Reddit
- Jewellery retailers quietly expanding lab-grown SKUs to meet demand they were already seeing
None of that showed up in a quarterly revenue report. All of it was visible in consumer sentiment shifts months (if not years) in advance, if anyone had been looking.
There is also an incentive problem baked into most reporting structures, and it’s worth naming directly. Lagging indicators are easy to defend in a leadership meeting: they are precise, they are comparable quarter over quarter, and nobody gets challenged for reporting a hard number.
Leading indicators are messier. They require judgment calls about thresholds, and they can be wrong in ways revenue rarely is. That asymmetry quietly drags most organisations toward the metrics that are easiest to present, not the metrics that are most useful for deciding what to do next month. Breaking that bias is a leadership choice, not a tooling problem.
Quantifying Intent & Price Elasticity
Talking about “leading indicators” in the abstract is easy. Building a system that quantifies them is the actual job. Three methodologies do the heavy lifting.
1. Conjoint analysis
Conjoint analysis models how consumers trade off attributes, price, quality, brand, ethics and origin when forced to choose between alternatives. Applied to diamonds, it answers the question De Beers needed answered years ago: at what price gap does a consumer switch from “natural” to “lab-grown,” given that the two are visually and structurally identical to the naked eye?
This is price elasticity modelling at its most practical. It does not only tell you that demand is price-sensitive; it tells you where the breaking point sits, category by category, segment by segment. Any brand facing a credible low-cost, functionally-equivalent alternative should be running this analysis continuously, not once a year.
2. Search intent metrics
Search behaviour is one of the most honest leading indicators available, because it captures intent before a purchase decision is made, often before the consumer has consciously decided anything.
Tracking the ratio of high-intent category terms over time, “natural diamond ring” against “lab grown diamond ring,” for instance, reveals share-of-consideration shifting in real time, well before it reveals itself in share-of-wallet. A brand monitoring this ratio quarterly would have caught the crossover point years before rough diamond prices cracked.
This is the practical, buildable core of a marketing analytics early warning system: a dashboard tracking category-level search intent ratios, refreshed continuously, with alert thresholds when the ratio moves outside historical bands.
3. NLP and sentiment tracking
Search tells you what people are looking for. Natural language processing on social and review data tells you why.
Sentiment tracking at scale can isolate which value drivers are gaining or losing ground: ethics and sourcing, sustainability, price-to-carat perception, emotional narrative, and quantify the rate of change. When “ethical sourcing” mentions climb relative to “rarity” mentions in a category’s social conversation, that’s a leading signal that your value proposition needs to evolve before your revenue tells you so.
Together, these three methods form the analytical backbone of predictive marketing analytics: not forecasting revenue from revenue, but forecasting revenue from behaviour.
What makes this combination powerful is that each method compensates for the others’ blind spots. Conjoint analysis tells you where consumers will land if forced to choose, but it’s a snapshot, not a trend line. Search intent metrics show you the trend line, but not the reasoning behind it.
Sentiment and NLP fill that gap by surfacing the “why” in consumers’ own language, often before they’ve made a purchase decision at all. Run in isolation, each method gives you a partial picture. Run together, on a recurring cadence, they give you something closer to a live read on where a category’s centre of gravity is moving, which is precisely the intelligence a revenue report cannot provide until the shift is already irreversible.
A Blueprint for Building Your Own Disruption Early Warning System
None of this requires a diamond-industry budget. It requires a deliberate framework and the discipline to review it as often as you review revenue. Here’s a practical four-step build.
Step 1: Monitor intent ratios continuously. Identify the 3–5 keyword pairs that best capture your category’s core tension (premium vs alternative, brand vs generic, your product vs a substitute that threatens it). Track the ratio weekly, not quarterly. Set an alert threshold; a 10–15% shift over a rolling 90-day window is usually meaningful.
Step 2: Mine qualitative social data at scale. Run NLP sentiment analysis across social platforms, review sites, and forums relevant to your category. Don’t just track volume or overall sentiment; track which value drivers are moving. A stable sentiment score can mask a major shift in why people feel that way.
Step 3: Track competitor and substitute price compression. Price compression among alternatives is often the earliest hard signal of margin risk. If a substitute’s price is falling faster than adoption is growing, that’s a warning that the substitute is scaling production or distribution ahead of demand, exactly what happened as lab-grown diamond manufacturing matured.
Step 4: Sync consideration metrics with supply chain and inventory data. This is the step most marketing teams skip because it requires cross-functional access. But consideration metrics mean little in isolation. When rising interest in an alternative coincides with retailers quietly expanding its inventory allocation, you have corroborating leading signals from two independent sources, which is a far stronger warning than either alone.
Build these four steps into a single dashboard, reviewed with the same cadence and seriousness as your revenue dashboard, and you have the foundation of a genuine market disruption analytics capability, not a nice-to-have report, but a functioning warning system.
One implementation note matters more than any tool choice: cadence beats sophistication. A simple dashboard, reviewed weekly by the people who can actually act on it, will outperform an elaborate model, reviewed quarterly by a committee. The diamond industry didn’t lack analytical talent. It lacked a forum where leading signals were reviewed with the urgency of an earnings call. Fix the review cadence first. Add sophistication after the habit is established, not before.
Read also: Kenyan Brand Benchmark Report 2026: Top Brands, Consumer Trends, and Market Insights
Conclusion
The diamond industry’s crisis isn’t really about diamonds. It’s about what happens when an organisation’s measurement systems are built entirely to explain the past, with nothing built to anticipate the future.
De Beers wasn’t short on data. It was short on the right data, refreshed at the right speed, feeding decisions at the right level of the organisation. That gap, not lab-grown diamonds themselves, is what turned a manageable trend into an existential crisis.
Three takeaways for your next analytics review:
Audit your dashboard for lag. If every metric on your executive dashboard is an outcome (revenue, churn, share) and none of them falls in the behaviour category (intent, sentiment, search ratios), you are flying backwards.
Give leading indicators a seat at the table, not a footnote in an appendix. If sentiment and intent data live in a separate report that nobody reviews with the same rigour as revenue, they aren’t actually informing decisions; they are a decoration.
Set alert thresholds, not just historical charts. A dashboard that shows trends without triggering action at defined thresholds is a museum piece, not a warning system.
The natural diamond industry had years of warning signs sitting in search data, social sentiment, and price elasticity models before the crisis hit its revenue line. The lesson isn’t that De Beers should have predicted the future; it’s that the future was already visible, in data most companies already have access to but rarely treat with the urgency it deserves.
Pull up your own analytics stack this week. Count how many of your leading metrics actually trigger a decision versus how many just decorate a report. That ratio, more than any revenue number, will tell you how exposed you really are.
At Marketing Analytics Africa (MAA), we are passionate about empowering African businesses with data-driven marketing strategies. By providing the tools and insights necessary to navigate this transformative journey, we help organisations unlock their full potential in today’s data-centric world.
Ready to unlock the power of data for your business? Let’s get the conversation started.
