Let’s say that you have a widget that you would like to sell for $100. This widget will earn an after-tax profit of $81 (impressive tax rate courtesy of NOLs), so you should be indifferent between these marketing options:
+ $80 buying 80,000 impressions at a $1 per thousand impressions (CPM) with a 0.25% click-rate and 0.50% conversion from click to purchase.
+ $80 for 200 clicks at a cost per click (CPC) of $0.40 each with a 0.50% conversion from click to purchase.
+ $80 fee per purchase / action (CPA).
This is a very simple analysis and assumes that each customer is a one-time buyer.
Building a business on $1 CPMs.
In the example above, we showed that a $1 CPM with 0.25% click-through rate and a 0.50% conversion to purchase from click would be the same as getting paid $80 for every purchase from our hypothetical advertiser. So if we assume that a $1 CPM is about right and figure out what audience is required to build a $100MM annual revenue business, we find out that we need 8.33 billion monthly page views and over 300 million monthly unique visitors creating 25 page views per user.
With these assumptions, we would need a site that ranked fourth globally, according to data from comScore. Of course we could assume that we get 10x the page views per user, but we’re still at 33 million unique visitors per month. It’s a non-trivial feat to develop a $100MM annual revenue business through online advertising.
What we can learn from Google about making money on the web.
Google creates about 10,000,000,000 paid clicks per quarter at a little over $0.50 per click for revenue of ~$5 billion per quarter. For a page view driven model to yield the same number of clicks at a 0.25% click-through-rate, they would need 4 trillion page views per quarter (10B / 0.25%)! According to Wall Street research I read recently, Yahoo! is under 400 billion page views per quarter.
Many sales and other internet executives inside CPM-dominated organizations will argue that brand is about so much more than clicks. Unfortunately, there isn’t a great way for those executives or any one else to objectively measure the intrinsic value of one impression over another. Measuring leads or purchases is quite simple, on the other hand.
The problem this presents for product development is that there is no objective function — there isn’t one thing that a team can optimize to other than impressions in this model. And that’s exactly what you frequently get — more impressions. Even if it means turning your outbound links inbound to keep a user at your site or forgoing a feature that decreases impressions (but may improve user experience). This harms consumer loyalty and long-term usage. It’s also hard to see how this increases value for brand advertisers.
I suspect that if product groups optimized to some other metric (e.g., clicks, purchases) that both the consumer and advertiser would benefit. Google has had the luxury of having extreme clarity in terms of having an objective that serves the consumer and advertiser almost equally well — and one that is extremely measurable in clicks (paid and unpaid).
How to make money on the consumer web.
We are in a recession. Worse, we are entering a prolonged period of weak consumer spending thanks to the housing debacle and the energy crises. Discretionary marketing dollars (those without a clear ROI) will not grow at the same rate as internet audience — this will lead to an aggregate decrease in brand-based CPMs over time.
If you want to build a real business in consumer, I would encourage you to either sell something (software like Microsoft or Adobe, hardware like Apple, some service like eFax, ecommerce like Ebay or Amazon) or figure out how to build a business that measures and delivers REAL VALUE to advertisers.
Even if you are a CPM advertiser today, I would encourage you to measure outbound clicks from your site/s as a measure of intrinsic value. You may not be charging by clicks today, but it cannot hurt you to understand the intrinsic lead generation capacity of every page of your site. The difference between direct marketing and brand marketing is one of timing and repetition. The lines between the two will fade away over time. And my money is with publishers who understand the lead generating capacity of their properties.
You can find a copy of the revenue.xls model from this post, here.
7 responses so far ↓
Peter Cranstone // September 14, 2008 at 4:41 pm |
Excellent post!
Everybody who has a business model that is “ad supported” should read it. As you say, it’s non trivial to build a $100m a year business on ads.
There are virtually no barriers to a web 2.0 business other than being first and the bloom is long off that rose.
Building a business with measurable, sustainable, profitable revenue from volume is really hard work and requires just the right opportunity and great execution.
Building a web site to 300 million monthly is a daunting prospect, especially when you consider the competition is exactly one click away.
Cheers,
Peter
Jon Bischke // September 14, 2008 at 7:46 pm |
Great post Mike. You hit on exactly what makes ad-supported businesses so challenging (the difficulty in determining the amount of true value creation). While online is actually at an advantage to offline in this respect, it’s still often tricky to figure out where value has been created. And if it’s not being created in the long run that will most certainly be exposed regardless of any hype, funding, impressive growth metrics that might used to substitute for value creation in the short term.
Sean O'Malley // September 15, 2008 at 8:05 pm |
“If you want to build a real business in consumer, I would encourage you to either sell something (software like Microsoft or Adobe, hardware like Apple, some service like eFax, ecommerce like Ebay or Amazon) or figure out how to build a business that measures and delivers REAL VALUE to advertisers.”
Mike,
you captured several concepts that I’ve been pushing for some time. Firstly, that ad supported businesses are difficult to build. More importantly, that any business needs to be centered around real value. And when it comes to new businesses on the web I’m more inclined to build something that has a price or transaction involved. BTW, I was just writing about this in my latest blog around Fitbit which you can find on my blog at seancomalley.com.
Peter Cranstone // September 15, 2008 at 10:14 pm |
Mike,
I have a follow up. What would be really interesting is knowing more about the expense (capital requirements) side of the house.
Let’s assume (always dangerous) that you have an incredible idea for a Web 2.0 play and it’s already growing leaps and bounds.
What does it cost you in equipment to scale to a business that does 8 billion page views a month. My guess is that over 5 years you would need at least $25m in capital and I would expect that to be just the beginning.
What are your thoughts?
Cheers,
Peter
Mike Speiser // September 17, 2008 at 3:42 am |
I’m not too worried about the operational expense of a web site. I can’t imagine that your marginal cost of serving one thousand pages would exceed the hypothetical $1.00 revenue per thousand pages. And firms like Rackspace and Amazon’s web services efforts help startups keep the fixed costs of operations relatively low.
Of course, if you need to crawl the web to build a search index before you know if anyone wants your product, you may need significant capital in advance of traffic which is a huge risk. But even with search, most of the hardware would likely be serving up results (rather than crawling and indexing) which would scale with usage.
The marginal cost of video is the only significant marginal site cost these days. While an order of magnitude higher than everything else, it’s not that bad and the trend is your friend on video (caching and bandwidth are getting cheaper as I type).
The bigger issue is betting that your product will attract monthly unique visitor usage on the scale of something between the population of Sweden and the United States. Risky bet. If you’re that good (or lucky), I’m not worried about the operating costs of the site…
Benjamin Kuo’s Blog » Blog Archive » The math behind Internet advertising businesses // September 17, 2008 at 4:20 pm |
[...] Speiser, a Managing Director at Sutter Hill Ventures, has a good post from this weekend doing the math behind ad-supported businesses. Worth a read — his argument, that at a $1 CPM, you would need to be the fourth ranked site [...]
Finding your web business model - techfounder // July 1, 2009 at 10:57 pm |
[...] fact, those who can actually make it stick are few and far between (Mike Speiser shows the numbers don't lie). The amount of traffic you need to generate in order to make substantial revenue from online [...]