Amid the rhetoric over how to support the business of journalism, the City University of New York’s Graduate School of Journalism has taken a plunge by creating models of what a news network on the Web might look like with hard, if optimistic, numbers.
Broad and multi-tiered spreadsheets filled with detail build upon each other until at the bottom, they predict handsome profit margins for small, medium and large blogs, as well as a larger network that might aggregate them and have its own editorial and business staff. In the main cases, the sites are supported by advertising supplemented with other income streams.
The models, presented first this summer at the Aspen Institute and again last week at the “New Business Models for News” conference at CUNY, are a fine start, and the spreadsheets are provided openly. At both conferences, project leader Jeff Jarvis of the CUNY J-school said he was hoping people would test the models to strengthen their viability. “The aim,” he said last week, “is to question, poke, probe, improve the business models.”
My poking and probing shows that when the assumptions are changed to less optimistic but still reasonable scenarios, the models can swing to much lower profit levels or even losses.
As a framework, the models’ creators — consultants Jeff Mignon and Nancy Wang of Mignon Media and Jennifer McFadden, who is with the business of news project at CUNY — assumed a metropolitan region like Boston’s, with 5 million people. They also assumed that the main newspaper had gone out of business.
In one model, they posited that the population would get its local news from dozens of “hyperlocal” blogs, each serving smaller communities of between 20,000 and 60,000 people. After three years, a typical “large” blog serving 60,000 users with a staff of four would have a profit of $51,502, for a “net margin” of about16 percent on revenues of $331,640. A medium-sized blog with a staff of “2.5” would have $24,851 in profit on revenues of $168,420; a small blog with two staffers would make $10,739 profit on $93,089 in revenue.
Another model envisioned a news organization serving the entire population of 5 million while “working collaboratively with other members of the ecosystem.” The organization is projected to post a $348,761 loss the first year, a $2,054,515 after-tax profit the second year for an 18 percent net margin and $5,886,842 after-tax profit in the third year, for a 29 percent margin on $20,301,862 of revenue in the third year. Before-tax earnings are $9,811,404.
Mignon and Wang assured me the assumptions are all reasonable based on real world figures and experience, and I agree each of them taken alone probably is. But it’s also quite possible that some of them won’t hold true.
Let’s tackle the news organization serving 5 million people to test the point. (If you’d like to follow along, you can download the spreadsheet and plug in the numbers. Find the correct tab at the bottom — for example, “Revenue-Website” or “Unique Visitor Calculations” — plug in the new numbers then look at the bottom line “earnings” figure on the “Income Statement” tab which is first on the left.)
In the “new, metro-wide news organization” model, 80 percent of the population has Web access and 24 percent of them will view the Web property an average of 12 times per month, zooming up to 73 percent going to the site that often by year three. The site gets an average CPM (cost per thousand) advertising page rate of $12, and there will be supplemental income through things like events, coupons and iPhone apps. (Though many at Aspen and the CUNY conference quibbled with the CPM rate, it strikes me as reasonable for a page with multiple ads; the model also assumes that up to 40 percent of the ad inventory remains unsold.)
If we shave the assumptions bit by bit, we start to see a strong effect on the bottom line. For example, is it reasonable to assume the average user will visit 12 times per month? In my experience, news site visits go as low as 2-4 times per month for the average visitor. Nielsen Online reports that Boston.com gets just over four visits per month per user, according to a story published Tuesday by Editor & Publisher. The story also reported that the number of unique visitors for many newspaper sites is going down dramatically.
If we change the number of visits per month to four per user (on the Unique Visitors tab), the model swings to a $103,206 loss in year two; in year three, it cuts the after-tax profit by more than half to $2.1 million. Mignon conceded that there’s also downward pressure on ad inventory pricing, so let’s knock the CPM in year three down to $10 (Revenue-Website tab). And can we really expect 73 percent of people with Internet access in the region to get news from that site? If we knock the level down to 40 percent (Unique Visitors tab) — also a reasonable assumption, and still considerable growth from the optimistic 24 percent in year one — the profit goes down to $1.2 million. (You can see a spreadsheet with my new inputs here.)
That profit also assumes there’s no debt — no bank loans or similar obligations to investors, and also no amortization or depreciation, all of which any CFO will tell you can have a big effect on the bottom line. Someone with deep pockets will certainly have to fund the venture since there’s a presumed $349,000 loss in the first year even under the original assumptions.
The expense side of the equation can also be challenged. For example, the hosting and technology costs stay flat at $100,000 per year (Expenses-Website tab) for all three years, but as traffic balloons, won’t costs go up as well? Would a qualified CFO or COO in an area like Boston be willing to work for the listed base salary of $78,000 (Staffing tab)? I may have missed it, but I don’t see a line for legal or other professional costs. You can imagine the net effect these assumptions, if changed, will have on the bottom line.
None of this is to say that the models aren’t useful. They are. They give a framework, show their projections — if not always explaining the precise reasons for relying on these numbers — and allow a business owner to plug in his or her true figures and get a sense of what may be possible. Someone can make their own assumptions about audience, ad revenue, subscription revenue and so on after research in their region.
Wang said the team hopes that people do plug their own numbers into these models. “We only put one stake in the ground,” she said. “Any little tweak in any of the assumptions can greatly change the bottom line. We are very careful that the assumptions we put in there are based on our experience, or our client organizations.”
I would also recommend that a would-be owner trying to get at the real value of such a project determine “best,” “worst,” and “likely” scenarios and weigh them to help make a “go” or “no go” decision. I might also want to project the future earnings back into the present using an average weighted cost of capital with risk factors included. (These are the methods many investors would use to figure out in present dollars how much the business is likely to be worth in the future, which will help determine whether they should put their money into the business today.)
If this is getting too financially geeky, don’t worry. Wang, who attended business school at both MIT and Columbia, acknowledged that the models could use more work. But, she said, “You have to start somewhere.” The team was looking to complete the project, funded by a Knight foundation grant, in just a few months, in time for the Aspen presentation.
Mignon, whom I’ve known for years, and I also discussed other issues, such as the trend of advertisers demanding more measurable ad results. More and more, advertisers are looking for things like “acquisitions” — sales directly attributed to the ads — and the number of sales leads generated, rather than the simple number of ads viewed sold on a CPM basis. “My main concern when we are speaking about that is advertising is clearly shifting from eyeballs to ROI,” Mignon said.
Other questions include how different geographies with fewer residents might change the model, and how any news blog can fend off competition when a successful neighbor encroaches into its territory. There’s also the question of whether advertisers who are increasingly able to create their own content for the Web will even want to see their ads on news sites — a topic I’ll take up in a later post.
None of this means a news blog can’t succeed, or that it shouldn’t try.
“It’s all about execution,” Mignon said. “Some people can succeed and do it, when 95 percent of the people are going to fail. But we don’t say restaurants are not a sustainable business just because one out of two of them doesn’t succeed.”
Indeed, making any business a success entails a lot more than manipulating a spreadsheet. People using the CUNY model need to do their own projections and calculations, construct a number of scenarios with varied costs and revenues, and adjust as experience proves the predictions right or wrong. As Wang said: It’s a start.
Disclosure: I was asked last Spring if I might want to work on CUNY’s project but we didn’t pursue it.