Nieman Reports published an article by Hanaa’ Tameez in early July that described the unconventional, “no-platforms” approach of Gazetteer SF, a four-month-old subscription news site in San Francisco. Among other things, this site:

  • Does not track its traffic via common web analytics suites.
  • Does not have official social media accounts.
  • Minimizes its presence in search engines.
  • Uses minimalist software-as-a-service for its content management, email newsletters and SMS alerts.

This almost analog approach to a digital news organization intrigues me. It deliberately eschews technologies and services that, for most peers, are de facto standards for everyday operations and measures of digital content organizations.

But why? Here’s what Byron Perry, founder and CEO, said in the article:

“It’s just universal that everybody who has worked at a traffic-driving publication, which is most publications, has extreme burnout and hated it ultimately. It’s certainly not sustainable for the writers. And so they all burn out, they quit, they move to another job. They might leave journalism. Maybe worst of all, it’s not even working for the publications commercially.”

I understand fully why journalists might prefer jumping off the social-posting, search-optimizing, numbers-focused treadmill. Pushing a digital content organization to drive traffic (especially emphasizing near-real-time metrics) can resemble a collective dopamine quest, affecting staffing, news judgment, user experience, writing and editing styles. The article provides examples of how relief from that chase helps the Gazetteer SF news team be more creative and thoughtful in its efforts.

But it was Perry’s last statement — though I know some media executives and service providers would disagree with it — that struck me as worth further exploration. How well do widely used, technology-driven “best practices” work for media businesses? At what scale? And is performance trending better or worse over time?

Let’s take a very common digital business activity — advertising — and examine it in this fashion. The major technologies that enable ad businesses include:

  • Digital ad serving/ad management platforms.
  • Programmatic ad networks and exchanges, both for backfill of unsold on-site inventory and for media outlets that sell off-site network inventory to their clients as extended reach.
  • “Passive” ad distribution networks.
  • Search ads, social media ads and the management platforms that support them, for media outlets that resell these ads to their clients.
  • And, of course, the same traffic metrics/analytics, social and search platforms that Gazetteer SF deemphasized.

Nothing I say here is intended to insult any of these platforms, technologies, networks or services that enable today’s digital ad ecosystem. Quite the opposite. The ecosystem features remarkable technologies and services that provide very good return on investment for media businesses that can capitalize on the current ways digital ads are bought, sold, delivered and measured.

My work at LMA/LMF focuses mostly on small to medium-sized local media outlets. Most of them have, at their core, a legacy distribution vehicle — newspaper or broadcast — and have been developing their digital businesses over, say, two decades at least. Pretty much all of them run digital advertising and marketing services businesses of various forms.

And many of them, especially at the smaller end of the scale, tell us they are struggling to grow digital advertising businesses and keep up with the current ecosystem’s enabling tech, which often seems optimized for much greater scale and network effects than they will ever achieve. We hear that large chunks of their on-site digital ad inventory (display, video ads, branded content teaser ads etc.) go unsold locally and are filled instead with programmatic network advertising.

We also hear their struggles with the economics and user experience of these nonlocal ads — low rates and declining revenues, ad fraud, lagging load times, clickbait and salacious or dubious ads that, to be polite, don’t exactly enhance the experience or brand reputation of a local news and information site.

I touched on these issues in an article I wrote in April, encouraging media leaders to see opportunities even when certain business scenarios may turn perilous. In one example, I imagined how a media organization might reinvent its digital ad portfolio if programmatic advertising simply went away (for example, as a dramatic consequence of the ongoing shift away from third-party cookies — which, it appears, just slowed substantially). My intended takeaway/catch phrase in that post: Stop depending on systems to sell our stuff for us.

Now I’m trying to imagine how a smallish local media organization, one that has attempted unsuccessfully to operate using the common contemporary model for digital ads, might rethink its ad model — regardless of whether ad tech or programmatic networks change profoundly.

LMA is all about reinventing business models for news, but this would be more uninventing the current digital ad model than reinventing — picking up cues from old-school media advertising.

Could the team running a small website just “switch off” its ad server, programmatic exchanges, passive ad placements, and traditional models that emphasize impressions and/or clicks? Would it make things better or worse? And by extension, could a digital startup or new niche product begin its advertising business this way, without ever pursuing current industry norms?

We can examine some hypotheses about running a digital ad business without major components of the current ecosystem. For example:

Could you run a digital ad business without an ad server?

Once upon a time, we produced websites without database-driven platforms like content management systems or ad servers. The earliest website display ads often appeared on pages thanks to simple, home-grown scripts — or even were hard-coded into the page markup — and relied on redirects to count clicks, if you counted them at all.

I know of no one doing this today and I can understand why not. But the path is still available, and might even be a bit easier to execute vs. the “olden days,” using no-code or low-code automations without having a web developer on staff. Or you could go lukewarm-turkey and still use an ad serving platform with minimal capabilities.

  • Upside: Potential cost savings. Simpler product model might be easier to sell locally.
  • Downside: Though you might scale down to minimum functionality, most outlets would still need some ad serving technology just to house and deliver video and display ads. The cost to hard-code or otherwise cobble a simple process together — plus tracking placements, updating creative etc. on more than a few campaigns — could quickly exceed whatever you saved by dropping your ad server subscription. You probably still need to count impressions, deliver simple ad rotations and manage shares of voice. Some agencies, even those that represent local or regional clients, require some minimum standard capabilities from media outlets.
  • The verdict: For any media outlet with more than a few ad clients and the simplest needs, this move might be taking a bare-bones approach too far.

Could you eliminate all forms of network backfill?

Dropping remnant ads and passive-revenue avails means you become the only gatekeeper for your inventory. No paid ads would appear on your digital properties from sources you didn’t know, but of course, revenue associated with remnant and passive-fill ads would go away.

Could you rethink how you display ad messages across the board?

I always advocate fewer, larger web display units on websites that carry ads, especially in the current mobile-first era of websites. As a companion to this idea, you could replace some web display “boxes” with more native image/head/teaser formats to promote branded content.

The “fewer/larger” approach creates better share-of-voice for each ad message relative to all the other content on the page, including other ads. It reduces reliance on tricks like “homepage takeovers,” video autoplays or interruptive pop-up-style ads, which do not play out well at all in phone-size browsers. Improved impact, in turn, somewhat mitigates loss of available impressions from the reduction in units — which itself should not be a problem to any site that today turns half or more of its inventory over to remnant.

  • Upside: Generally better looking, better organized, better performing web pages, with less of a “minor-league outfield wall” appearance.
  • Downside: This change cuts the number of total display impressions you have to sell or otherwise monetize. It may also take you farther away from industry standard unit sizes, though you might find more than enough to work with using contemporary fixed-dimension standards. Even so, agency media buyers and planners may gravitate to older, more common unit dimensions.
  • The verdict: It’s worth a try if your display inventory is nowhere near sold out, and your digital business doesn’t depend too much on agency buyers who still obsess over the common leaderboard, smartphone banner and skyscraper units.

Could you go (or go back) to share-of-voice instead of impressions?

Smaller and or newer sites with too little traffic to justify impression-based selling could consider an old-school way. Instead of impressions measured in the thousands to hundreds of thousands, percentage shares of the total inventory become the atomic unit of digital ad sales. In such a share-of-voice model, ad campaigns would be delivered across the site’s inventory (or subsets of it) in random rotations allotting each campaign its predetermined share.

My longtime colleague, David Buonfiglio, used to champion a simple share-of-voice model for any new product launches where initial traffic was difficult to forecast. Example: Launching a new niche-content app for iOS and Android? Start by selling four clients equal 25% shares of whatever impressions you get, random rotation, no minimum guarantees — but with a minimum six-month commitment. This way, they’re “pilot partners” who share in launch promotion and are willing to gamble six months’ (or more) spend on future upside in traffic.

And by the way, another former colleague, Guy Tasaka, often reminds industry friends that native apps might not generate the same gross tonnage of unique visitors that media websites do — but the people who go to the trouble to download and use your native apps tend to be among your most loyal audiences, with much higher views-per-session and sessions-per-day ratios vs. website visitors. Apps see far fewer “one-and-done” user sessions. So early-adopter clients who buy into share-of-voice on a new native app launch might get smaller reach but greater frequency of visits by a predominantly local audience.

Besides new product launches or native apps, existing low-traffic, small-market sites that have failed to sell locally based on impressions might also consider such a change. If you don’t like handing over all that unsold inventory to remnant, you could revert to share-of-voice and figure out how to divide your whole inventory into shares to get and keep enough direct-sold clients on board.

  • Upside: Even low-traffic sites or apps could sell inventory on this basis. For new product launches, it gives advertisers the sense of getting in on the ground floor, with a calculated and manageable risk. Share-of-voice mostly eliminates the need for make-goods should traffic targets not be met.
  • Downside: If a product launch fails or traffic doesn’t at least come close to expectations, share-of-voice ad clients might quickly tire of waiting for better results. The SOV model also doesn’t work well if clients want much in the way of targeting or segmentation. Compelled by the need to measure everything, few agency buyers or media planners even think in these terms, let alone place campaigns this way. A site that brings in substantial agency business might be able to do SOV only in limited environments and time frames, such as the launch window for a new niche app. Finally, you still have to measure impressions and effective revenue per thousand — if your share sales can’t produce more revenue per impression than you get from remnant, this idea isn’t for you.
  • The verdict: It’s worth a try if traffic is modest and you are not well invested or successful in impression-based sales.

Who’s experimenting? Who wants to?

Gazetteer SF’s platform-minimizing approach fits its content model (don’t let traffic metrics sway news judgment) and even its business model (a hard paywall for subscriptions). The approach sounds good for that particular startup, but the story might be different if it launched with a heavy dependence on advertising.

Because decisions about alternative ad models affect revenue, real and potential, one must be careful not to jump to conclusions based on what sounds good. I know far too little to tell anyone definitely do this or definitely don’t.

My instincts, however, say any of these alternatives, alone or combined, might work better than the current models for digital advertising at small scale and under specific conditions.

Because so many of us in the media industry have long relied on the current models — and so much of the digital ad ecosystem is built to enable them — we may find even experimentation with alternatives is too challenging.

Still, we have so many bright, inquisitive people out there running smaller digital media businesses. If you have tried, are trying or want to try anything resembling the ideas discussed here, I would like to hear about it. Please fill me in.