Yesterday, Amazon founder and CEO Jeff Bezos bought The Washington Post, effectively ending the Graham family’s 4-generation long stewardship of one of the country’s most accomplished newspapers. Bezos’s purchase matters to journalism because it suggests a new era in the meaning of media ownership — one that requires tracing influence through infrastructure design.

First, a few distinctions: Bezos, not Amazon, purchased The Post; the paper will not be a subsidiary of Amazon or share a parent company. The paper’s editorial leadership remains, with Katharine Weymouth (niece of Post Chairman Don Graham) staying on as Publisher and CEO. But beyond these points, it’s unclear exactly what the purchase means.

Which makes this precisely the right moment to ask: going forward, how might we trace what the purchase could mean?

Ownership, control and the news

It’s incredibly difficult to study how ownership influences the media, and why such influence might matter. Sometimes media ownership matters because there’s evidence that too much media power in too few hands is bad for democracy: concentrated ownership doesn’t produce the kind of diversity self-government needs; it makes it less likely that powerful people will be held accountable; and it prevents markets from self-adjusting and reallocating resources to newcomers.

Other times, ownership is a way to see organizational priorities. Owners and publishers get to exert what Murdock calls “allocative control”: influencing policy and strategy, merger and acquisition decisions, profit distributions, shareholder communications, or labor allocations. These moves are distinct from the day-to-day “operational control” that individual editors and reporters have as they decide what issues to cover, what headlines to write, which sources to interview, or where to place a story.

At other times, ownership is thought to be about more blatant ideological influence — elite media owners imposing their values and priorities on their journalistic employees, if not in editorial page statements, ensuring that news coverage reflects how they see the world and how they think it should change. Even when news organizations aren’t financially profitable, they might still give their owners valuable channels for political influence.

And, finally, media ownership can correlate with self-censorship. Even without explicit instructions, journalists might simply sense — through the tacit social cues and agreed-upon patterns that often structure newsrooms — that their time and attention would simply be better spent elsewhere, not on issues that are implicitly understood to be out of alignment with what their organization’s culture or leadership is thought to value.

Bezos’s purchase of The Post, though, made me think not of these types of control but of more subtle and less visible dynamics of control that Daniel Chomsky (not that Chomsky) found at The New York Times in the 1950s and 60s. By systematically combing through an archive of correspondence between Times editor Turner Catledge and Times publisher Arthur Hays Sulzberger, Chomsky observed the following pattern: Sulzberger would write to Catledge with “suggestions, instructions, complaints, and orders” (at least several hundred times over a 7-year period); Catledge would then strip these memos of identifying information and forward them to his employees as feedback from an “interested reader” (suggestions from Sulzberger’s wife were described as being from “a friend of an interested reader”). In all but two cases, Catledge ensured that Sulzberger’s wishes were reflected in paper’s coverage, letting reporters be under the impression that they were simply receiving direction from their editor. Their sense of professional autonomy from their owner’s control remained intact, Catledge retained his colleagues’ respect as an independent editor, and Sulzberger enjoyed anonymous control over the paper.

This kind of influence is probably rare and most reasonably seen as a function of the Sulzberger-Catledge relationship. But it has two notable features that bear relevance today: the true nature of the influence was invisible (Catledge was a willing accomplice in protecting Sulzberg’s anonymity) and it was enacted through organizational structures that let reporters retain a false sense of autonomy (they didn’t even know that they were being influenced by their owner).

It’s the relationship of these two things—invisibility and organizational structure—to professional autonomy that made me pause when I heard about Bezos’s purchase of The Post. What does this sale mean for understanding contemporary owners’ influences on online news organizations?

The answer, I think, lies in seeing influence in infrastructure design.

Invisible infrastructure and ownership influence

Amazon is infrastructure. It isn’t just an online retailer or a technology maker, but rather, one of the principal creators of the digital “scaffolding in the conduct of modern life”. Star & Ruhleder define infrastructure, in part, as: transparent (you usually don’t notice it until it fails); embedded within other technologies (it can’t exist completely on its own); standardized (dependent upon predictable behavior); and wide-reaching (it’s often surprising just how many dimensions of life a seemingly narrow infrastructure can influence). Recall Senator Joe Lieberman’s influence that made Amazon.com stop hosting WikiLeaks after the website carried classified information that had been leaked. It’s a mistake to see Amazon as simply a recommendation engine or a just-in-time shipper (although it’s those, too). Instead it should be seen as a maker and regulator of largely invisible infrastructure, which helps us see how the organization exerts “logistical power” to influence the social and technological conditions under which information can circulate.

This is how Bezos’s ownership might be analyzed. It seems highly unlikely that he’s going to exert editorial control over Weymouth in the way that Sulzberger influenced Catledge — or even that he’s going to pull a Post story that the government doesn’t like. Indeed, Weymouth retains her title as Publisher, in addition to Post CEO. Bezos explicitly states that The Post’s “duty will remain to its readers and not to the private interests of its owners,” and there’s no reason to doubt the sincerity of his statement.

What’s less likely, though, is that Bezos will curb the design thinking and infrastructure building instincts that helped him make Amazon a phenomenal internet success. The place to potentially see Bezos’s influence is in how he aims to reinvent “almost every element of the news business.” Indeed, it seems that The Post selected Bezos as its buyer for his skills and experience as an internet entrepreneur and networked system builder. As Weymouth puts it, “Mr. Bezos knows as well as anyone the opportunities that come with revolutionary technology when we understand how to make the most of it. Under his ownership and with his management savvy, we will be able to accelerate the pace and quality of innovation.” Don Graham seemed to agree as he described how he had reached his leadership limit: “the newspaper business continued to bring up questions to which we have no answers.” The Post needs new infrastructure.

I’ve previously described the infrastructure underpinning networked news production as “newsware”: the “networked technologies, algorithms, interfaces, practices, and norms that constitute the shared, embedded and largely invisible set of material and ideological conditions and logics governing press-public interactions online.” Every time a news organization relies on Facebook News Feed algorithms to surface its stories, on third-party widgets to organize its “most popular” reports, on Google News to index its content, or on Twitter to teach its journalists social media best practices, it must reinterpret the meaning of press freedom. It must find ways to define and assert its identity and values through infrastructure — infrastructures created and managed by new media organizations with very different histories that those of the family-owned Post.

To be clear, I don’t see Bezos’s purchase of The Post as some conspiratorial step in Amazonian expansion, nor as a way for Bezos to personally steer The Post’s journalism. This isn’t that kind of ownership move. Rather, my hunch is that it represents a moment in a new era of understanding what ownership and influence means for news organizations.

If we are to see Bezos’s influence over The Post, it’s probably going to come from tracing how his understanding of successful internet entrepreneurship intersect with the aims of democratic press. This overlap exists not in him directing The Post to cover particular issues or interview specific sources that align with his personal priorities (the way Sulzberger influenced Catledge at The Times). Rather, we might more accurately see his influence in The Post’s next generation of networked systems that control the role audiences play in news production, that use automation, relational databases, recommendation engines, and user interfaces to create online news experiences that “work” for audiences. Indeed, Bezos has already set a tone by saying that The Post’s “touchstone will be readers, understanding what they care about.” This is a laudable goal, and an important element of any news business that wants to build an audience and earn revenue — but the kinds of systems that deliver stories readers care about — that earn clicks, recommendations, likes, shares, and advertising revenue — are not necessarily the same systems that produce stories in the public interest.

What would it mean to recast media ownership concerns in terms of infrastructure design? What kind of design market produces sufficiently diverse newsware? What’s the difference between “allocative” and “operational” control when designing newsware — e.g., between systems that make it likely for news to circulate versus those that give individuals the power to share particular stories? Might news technology designers self-censor just as journalists sometimes have, creating systems that align with what social media signals and big data sets suggest readers will likely click on and share?

Tracing these kinds of infrastructural influence is different, but no less important, than uncovering the kind of control that Chomsky’s found at play at The Times in the 1960s. But we can’t afford to wait fifty years for the evidence to surface.

Image by ptufts used through a Creative Commons license.