The Dark Sides of Digital Marketing

Ghosts in the Dark: How to Prepare for Times of Hyper-Privacy

Felipe Thomaz


Dark Web, Privacy, Privacy Calculus, Privacy Paradox, Personalization

download pdf

A hidden digital romping ground is thriving in the dark
The Web is a dynamic, complex and rapidly evolving environment, with literal fortunes made and lost as the waves of change give rise to different business models. However, this complexity is deeper than immediately obvious. Apart from the surface web that most of us experience every day, there is a separate, hidden layer called the dark web. Here the websites are unindexed, access is only possible via Tor, a specialized browser, and communications are usually encrypted. Every aspect of the dark web is designed to provide privacy to its users. If you’ve heard of the dark web before, then it might be related to the number of illicit businesses operating there. It is the home of hackers, drug markets, data brokers, and human traffickers. However, it also serves as a safe haven for whistleblowers, activists, and journalists, as well as citizens from countries where comms are either restricted or monitored. Altogether, it is a place built for individuals who are exceptionally incentivized to be digitally invisible.

The dark web – a dorado for privacy
However, the dark web has its bright sides because it also exists as an unregulated testbed for technologies that we will eventually experience on the surface:  WhatsApp, as an example, provides similar end-to-end encrypted communication, and surface-web consumers start  gaining experience with bitcoin whereas in the dark web, they use it for a shadow economy with a GDP larger than Peru’s along with several other cryptocurrencies. That provides us with a useful place to study consumer privacy and have a view of what the surface world might look under an extreme level of consumer data protection. And the extent of that protection is astonishing. When I investigated my own digital footprint on the surface web, I found close to 5000 tracked variables across data aggregators, ranging from absurd guesses to oddly specific details that have been accumulated over the past decade. Conversely, the average ‘persistent’ dark web user  – one that decides to keep the same persona for more than a single burst of use – had eight data points. Yet, most smart dark web users who forego this persistent personal branding could choose to leave essentially zero trace of their existence. And with that, they become invisible. Ghosts.

The dark web’s privacy – a nightmare for marketers
This reality should be terrifying to anyone relying on the modern marketing machinery that fuels much of today’s business growth and competitive edge, as all of it relies on abundant information. Lookalike matching, collaborative filtering, precision targeting, audience controls; all of it disappears if everyone is a ghost. Even our best customers will look like never-before-seen individuals until they decide to reveal themselves, by logging in or entering ID information, for instance. And as a result, marketers would be reduced to pre-information-age tactics, reliant on population averages, and at best using unsupervised machine learning techniques, like clustering.

The privacy calculus: Ghosts or Buffs?
But not everything is ghosts, doom and gloom. Ghosting is a consumer choice. This choice between privacy and disclosure is called the privacy calculus. If there is trust, and a worthwhile value exchange, consumers might be willing to share their data and not enact all of the hyper-privacy available to them, continuing to give marketers full view of their behaviors and preferences. We call these customers ‘Buffs` (see Figure 1). For Buffs, marketers will have the full modern array of marketing and predictive analytics available, and provided they are doing their jobs well, one can imagine that this can only lead to higher profitability and retention rates.


Nudging consumers against the ghosting-option to share behavioral information and preferences
To explain how consumers can be convinced to be less secretive I like to use an analogy. The exercise is simple: first imagine your ideal lover/partner. This person knows you inside and out. Somehow, they always say the right thing, and intuit exactly what you need, when you need it. That is one set of behaviors, and one type of relationship. Now consider an alternative: this person has gone through your rubbish bin to try to figure out what you’ve been up to. They opened your mail behind your back, and once showed up uninvited to your parent’s house. They know where you are right now, and likely where you’re going next.

That’s probably enough to give a sense of these two relationships; and based on experience I can guess that we moved from “where do I find this magical person” to “someone call the police.” But for us, it is important to look at the details and ask why. The first person stayed within the boundaries of the relationship, likely accruing information little by little and over many interactions where they were expected to learn this information. And all of this was accomplished in the process of providing something of value. The other person; however, completely ignores normal relational boundaries and norms. They know things that they shouldn’t and act on that knowledge. This is the crazy lover/partner, who should be rightfully feared. Yet, we might allow our brands to become the crazy lover in the pursuit of profit, and in the fear in missing out – but for how long? In a world where consumers can opt out of the buff-position, a caring relationship is much more promising (see Box 1).


Will the hyper-private web become reality soon?
Thankfully, the hyper-private web will not exist, or is decades away, right? Well, that depends entirely on three groups: customers, legislators, and companies (see Figure 2). Customers could bring about a hyper-private surface web very quickly by adopting different behaviors and using specific technology. However, individuals seem unwilling to modify their digital behaviors or deploy new tech en masse. Furthermore, we also know that individuals claim to want more privacy, but still share information freely when asked, which is called the privacy paradox. So, ghosting will most likely not become a mass-phenomenon too quickly. The second group, legislators, appear highly motivated to enshrine privacy in regulation; however, their processes are slow, lack technical knowhow, and are enforceable only within their national boundaries. Altogether, a poor combination for meaningful change.


That leaves companies themselves as the most likely group to bring about hyper-privacy, and the reasoning behind it is strategic. Large, incumbent companies who already own significant amounts of data, and who have established strong customer relationships are exceedingly incentivized to create a hyper-private environment where new, challenger brands simply cannot easily and cheaply generate valuable data assets. Small losses of data to incumbents translate into massive losses to challengers, and privacy creates a significant barrier to entry. Anecdotally, significant players like Google, Apple, and Facebook have recently positioned themselves as privacy-first companies, sometimes losing advertising revenue to make it happen. Moves that are pro-consumer, but also highly profitable.

Manage customer choice and go for a meaningful “share of data”
So, it appears that we have a rapidly approaching dark-surface web that is hyper-private and full of ghosts. Marketing managers should therefore expect the access to data to decrease by default. Everyone will need to work with less, both in primary and secondary data sources. And with the lower data availability, the cost of acquiring data from suppliers and partner will increase.

If the ghost/buff position is a consumer choice, then companies must worry not only about share of market, but also their share of meaningful and actionable first-party data. Just imagine that you’ve lost your CRM data assets, cannot replace them, but your competitor is somehow intact. How long until they have an overwhelming competitive advantage? To seize the opportunities, companies should take stock of their customer-relationships, precise their data needs and learn what information is critical, advantageous, or irrelevant for their context. They should ensure that their brand is not the ‘crazy lover’ and implement initiatives that drive choice carefully in a trustful relationship like highlighted in Box 1. The“buff”-version of service that runs on full information will be more functional than a restricted, less-smart version for “ghosts”. The difference in functionality will serve as an incentive for customers to move from ghost to life-long buff.


Felipe Thomaz, Associate Professor of Marketing, Saïd Business School, University of Oxford, England, Felipe.Thomaz@sbs.ox.ac.uk

Further Reading

Thomaz, F.; Salge, C.; Karahanna, E.; & Hulland, J. (2020):” Learning from the Dark Web: leveraging conversational agents in the era of hyper-privacy to enhance marketing”, Journal of the Academy of Marketing Science, Vo. 48 (1), 43-63.