Lenin once said: “There are decades where nothing happens; and there are weeks where decades happen.”

Lenin’s politics aside, he said this in the early 20th century, before the internet, before algorithms. Today, change is not only happening faster, but everything is interconnected.  Economic, political, and cultural change move in tandem – creating overlapping disruption in what some like Adam Tooze have termed a polycrisis. Being relevant and having a deep understanding of context have never been more vital.

For brands, this is a reckoning – consumer priorities, loyalties, and the very tone of effective communication are evolving faster than ever before. Advertising that resonated pre-2020 can seem tone-deaf today; humour that worked last year may not land this year.

Yet many in the research and marketing industry still rely on historical norms – benchmarks built on attitudes and behaviours that are five or even 10 years old. This risks brand resonance and can even lead to backlash.

We saw this during the early days of the pandemic, when simple gestures like handshakes and hugs suddenly became taboo – forcing brands like Hershey’s and Geico to pull new creative featuring too much human contact. What was once heartwarming quickly became tone-deaf, highlighting just how fast consumer expectations can shift.

The industry needs living benchmarks

Norms aren’t just numbers – they reflect shared values and expectations across cultures, communities, and generations. They offer a snapshot of what ‘good’ looks like through comparative metrics like ad recall, brand lift, and purchase intent. 

There’s a tension here which is rarely acknowledged: the mathematical definition of a norm gives a business confidence; however, the cultural interpretation should lead us to question normative benchmarks that are overly reliant on historical data. Is statistical stability a misguided comfort blanket for our industry? 

Many research businesses are built on massive normative databases – extensive, longstanding collections of campaign performance data. Their sheer size gives them authority, offering reassurance to advertisers making decisions under pressure. However, the larger the dataset, the slower it becomes to adapt, as new trends struggle to move the weight of historical averages. Norms built on history can become a liability when they no longer reflect current realities. Compare this with streaming platforms and e-commerce businesses, which regularly recalibrate their algorithms to reflect changing consumer behaviours and capitalise on trends. You get diametrically opposed systems: one adaptive and one reluctant to change. 

Too often, norms live on as static benchmarks inside reports – rarely interrogated or refreshed. Dashboards may have democratised data, but when disconnected from consumer insight teams – the people trained to ask why – we risk mistaking legacy patterns for present-day truths. We take comfort in scale, without asking whether the data still reflects today’s consumer reality. In doing so, we risk optimising for a world that no longer exists.

Benchmarks should be a living representation of today’s consumer. That means actively managing datasets to ensure they reflect the present, not the past. It’s a similar principle to how streaming platforms recalibrate their algorithms in real time. As consumers, we don’t want recommendations based on our interests last year; we want our interests now reflected.

Living benchmarks are the foundation of connected data, giving brands the context they need to make decisions in tune with people and culture.

From living benchmarks to scalable insight

Of course, constantly refreshing benchmarks through new fieldwork can be expensive and slow – especially where a category level benchmark is required – for example, brands operating in niche verticals or emerging markets. That’s where AI and machine learning come in. Once you have a foundation of living benchmarks, you can use them to model new ones more quickly.

We’ve experimented with that, using our normative datasets to build synthetic benchmarks. By modelling country- and category-level effects, we created predictive norms where traditional data is limited or unavailable. Early results have shown up to 90% accuracy compared to a hold out set. While this work is at an early stage – it is a technical upgrade and by no means a commercial offering – and will no doubt prompt questions around synthetic data, it points to the offer of a way forward. It’s true that there is a lot of work to do to address questions around derived benchmarks, but the opportunity is huge.

For those not ready to take the leap to synthetic data, at the very least we should be scrutinising benchmark data. Perhaps we all need to ask in what instances are benchmarks holding us back as an industry rather than moving us forward – and adapt accordingly.

Evolving to meet consumers in stride

The biggest barrier to adopting synthetic norms isn't technical, it’s attitudinal and behavioural. There’s often a legacy in research teams – a distrust of so-called ‘black box’ systems which are seen as opaque and hard to trust. The industry has long relied on norms grounded in familiar, historically-observed data, providing comfort in their transparency and consistency. But comfort breeds complacency. 

The concept of dynamic, culturally relevant norms is especially compelling today, as consumer behaviour and expectations continue to shift rapidly. The question is clear: are your benchmarks helping you win with today’s consumers, or are they holding you back with yesterday’s data?

Jack Millership is head of data quality at Zappi