In just a few short years, e-commerce startup Fab went from a $1 billion valuation to a $15 million sale.

Across industries, success is more unpredictable than ever. When it comes to cultural products, things that worked in the past often do not work in the present (the sheer number of Avengers sequels notwithstanding). But despite the inherent unpredictability of consumer tastes and the complex way they interact, VCs still put a heavy bet on pattern recognition. These patterns — be it a proprietary product, low-cost customer acquisition tactics, or the ability to reach scale fast — are hardly reliable predictors of success.

For example, Harry’s proprietary product, manufactured in its German factory, was a great initial way to differentiate from Gillette’s low-quality but expensive razors. But superior product quality has since become table stakes in the shaving market, with a number of startups all offering the same key features. Five years and 375 million VC dollars later, Harry’s has only 5% market share in the traditional retail sales market, and is a distant third in the online manual shave market. Not until Walmart (ironically, the retailer that Harry’s DTC model set out to disrupt) provided its massive distribution muscle did Harry’s business start to shift. To stay competitive in this mass market, Harry’s now needs to worry about shelf space and brand marketing — just like Gillette.

Dollar Shave Club, with 21% of the online market share, was not profitable when Unilever bought it in 2016. Its VC-beloved debut online video was viewed more than 25 million times since 2012. Social media quickly made a lot of people know of Dollar Shave Club but also undid its staying power. The main lesson is that wide awareness doesn’t mean conversion and that seeing the fast user growth doesn’t mean profitability. To hack growth, startups have to hack culture first.

In addition to the usual signals, VCs should determine whether a company has roots in a subculture or trend. A subculture is made up of people who are more informed and passionate about a topic than anyone else. They are likely to be beta-testers, source material, and advocates for a new product or service. Cycling brand Rapha started from cycling obsessives. Apparel brand Patagonia started from the subculture of social responsibility. Deep subculture entrenchment ensures that a company can maintain and enhance its difference as it scales. Long-term defensibility has more to do with whether a company can believably connect with a community through the shared things they like than if it has a proprietary product or acquisition channels.

Success also has to do with what Japanese call kuuki wo yomu, or “reading the atmosphere.” In the October 2013 article titled “Yes, Real Men Drink Beer and Use Skin Moisturizer,” Bloomberg quotes Mintel’s data on the 5-year rise in the global sales of personal-care merchandise geared to men. Harry’s was founded earlier that year, and Dollar Shave Club two years prior. Both of them capitalized on the shift in the culture of modern masculinity, but neither of them invented it.

The shift was already happening. As sociologist Duncan Watts notes in his research on social influence, if a society is ready to embrace a trend, almost anyone can start one — and if it isn’t ready, then almost no one can. The success of Harry’s or Dollar Shave Club didn’t have to do much with a spiffy video or on the German factory-produced razors. It had more to do with how susceptible men already were to the idea of grooming and how easily persuaded they were to invest in it.

Social influence is often mistaken for disruption.

As the dynamics of how trends spread shifts from brands, media, and retailers pushing ideas to mass market, to the internet networks of niches and taste communities, both startups and VCs have to consider social processes that ultimately define the success of their inventions.

In addition to engineering products and services, startups need to engineer social influence in their market. The fastest way is to piggyback on already-existing social influence, and amplify it through a go-to-market strategy that emphasizes social activity among a company’s initial following. This social activity then serves as an ad for a product or service aimed at the mass audience. For example, luggage brand Away’s initial community of travelers — and their stories — became an ad for its products; rides of the Rapha’s Cycling Clubs are the ad for Rapha’s gear.

Social activity in a market accumulates social capital. How a social currency is going to be created and exchanged should be an inherent part of business plan. It’s a business’ core value unit, and whether a company has the potential to build and trade in social currency should become part of VCs evaluative criteria. Beauty brand Glossier’s currency is the beauty preferences of its fans. Glossier’s currency is so strong that this brand is now creating the entire marketplace around it. Social currency builds scale, defensibility, and network effects.

To prevent social currency from being devalued due to reverse network effects, companies need to maintain and grow their distinction as they scale. The best way to do this is through product and service diversification. A brand is an umbrella for a portfolio of unique products. The streetwear brand Supreme mastered the art of distinction, with a large part of its audience owning unique brand products and a limited number of people who own the exact same thing. Product diversification increases the number of bets, reduces risk, preserves social currency, and organizes a company around the inherent unpredictability of people’s tastes.

The ultimate irony of popular disruption narratives is that they venerate a deeply anti-social attitude. They celebrate an outsider and a renegade who “moves fast and breaks things.” But without social influence that creates the susceptible mood and allows new products, services, and ideas to spread, there is no “disruption.” Instead of applauding the world’s outliers, we should direct our attention to the society that makes them thrive. There should be a sociologist among the engineers.