The Official Blog of Cogmap, the Org Chart Wiki


Archive for July, 2008


OPA Study Reveals People Pay Attention to OPA

Thursday, July 31st, 2008

jim spanfeller

The Online Publishers Association (OPA) revealed a study that showed that members of the OPA have better performing inventory than other people.

True story.

K, that was kind of mean, but it felt a little Onion-esque to me so I had to roll with it. In particular, they single out portals and ad networks for their righteous slaughter. As always, they don’t give us an excel file to download to do our own analysis of the data, so I have a bunch of questions and problems. I won’t drill down into the “interactive & rich media” ad analysis they did, because my points are applicable there as well (in fact, more true!), but you can read this and then read the study and see all the same issues

“Ads on Content Sites provide double the brand favorability and purchase intent than advertising placed with ad networks.”

  1. So it seems like Dynamic Logic just aggregated every ad network study they have ever done here. Can we assume that we are basically talking about something that is maybe, MAYBE a slight premiumfrom a price/inventory mix to a RON ad network buy?
  2. So if ad networks ran 252 campaigns for the sample, and OPA members ran 1,185 campaigns, could one assume that OPA inventory tends to be viewed as more brand friendly inventory already?
  3. MOST IMPORTANTLY, so if we assume that the average ad network buy ended up being some slight premium to RON, maybe $1.75? (Really good networks might charge more for RON, but this is an aggregation of all networks, which means it is a lot of inexpensive UGC & Right Media), then is the inventory the OPA was selling sold for less than $3.50?

I have heard, and this is just hearsay, that apparently it costs more than $3.50 to buy inventory on ESPN,, CNN, and Cnet.

So even if the results of this study are true, if the advertiser has to pay more than $3.50 then the buy is inefficient. The advertiser is over-paying for the performance lift.

Fascinatingly, when PaidContent covered the OPA press release, Jim Spanfeller, the CEO of jumped into the comments and said (and didn’t say):

“…This study shows clearly that the premium environments offer substancail value above broad based non premium plans…”

A few things jump out:

  1. As a member of the OPA, his comments are in-line with the assumptions (including the flaws) that the OPA makes. The value delivered by premium environments is not necessarily in line with the cost to acquire the inventory.
  2. Despite his OPA membership, he does not talk about what kinds of inventory comprise a “premium environment” and that is probably because is starting their own ad network. There is no doubt in my mind that Forbes positions their network as “a premium environment” despite the fact that very little of their non-Forbes inventory probably comes from OPA publishers.
  3. So one thing I took away, implicitly, and he would probably never admit publicly, is that he may secretly believe that a well-targeted ad network campaign can perform just as well as premium inventory. Which kind of makes sense given the outcomes of this study. Could an ad network campaign with social media inventory perform twice as good as these RON studies? For sure.

Funny stuff.

(As always, these are my opinions, not my employers)

How Do We Build A Better Ad Business?

Thursday, July 31st, 2008

Innovators Dilemma

In an Ad Age article today, I once again read something Rob Norman said that I wanted to disagree with:

“The biggest gating factor to internet ad growth is the obsession of the players, the [venture capitalists] and the press with ‘bottom of the funnel’ marketing in a world where the big money is spent at the top,” said Rob Norman, CEO of Group M Interaction. He said display ads do indeed have a chance to be as big a business as the web giants hope — but it’ll require better targeting that helps discern intent (à la search advertising), a set of metrics that can at least be proxy for sales, and messaging that is valid and useful for users.

The second part of this sentence I completely agree with, but the means to the end I am less confident about. The first time I read this paragraph, I immediately thought of the Innovator’s Dilemma. It sounds like a classic example of a new technology targeting the “less valuable customers” with a radical, economics-redefining offering, perfecting it slowly, and gradually moving up market to take share.

Rather than write a ten page essay, let’s just agree that it sounds possible.  I actually like Rob.  He puts himself out there.  We would probably have a lot of fun hanging out.

Joining the conversation on advertising securitization

Tuesday, July 29th, 2008

Stolen From NYT.  Don't be mean.

I don’t think I have written my monster rant on aggregating conversations, apart from my rants about trackbacks. Regardless, lets all just agree that it is not good. Unfortunately, there are so many proposed solutions: Intense Debate, MyBlogLog, Disqus, FriendFeed, etc.., that it currently feels like the market is segmenting more as people search for the optimal standard. Healthy, but painful in the short term.

Anyway, the point is I left an interesting comment on an interesting blog post about an interesting article.

Go get it.

Stylehive Redux – Buried by Web 2.1

Thursday, July 24th, 2008

Well, Stylehive figured out a way to get me to come back: Send me spam.

Here is an email I got from them this morning and almost spam filtered:

Subject: Help me pick a hot bikini!

From: “” <>

Hey, Brent!

I’m going to the beach next weekend with the new boy, and I need help deciding what bikini I should get to surprise him with. I thought you could help me decide. Hive the ones you think are the hottest! Thanks in advance, I will let you know how this little experiment turns out.

To see all your messages and respond here

If you want to follow me just click here.

Happy Hiving!


I felt like this reeked of desperation so I went back to see what the hell was going on. So honeybee is a stylehive employee and she “used” the “send this member a message” button on thousands of members. How do I know she sent virtually everyone on Stylehive a message? Because she is getting dozens of new people “following” her this morning – more than 300 new friends this week.

Why does Stylehive need to send me sketchy email to check out their site? Probably traffic is flat. Let’s look:

Not exactly news, but traffic has flattened out. If they were getting $10 cpm’s per page, an outrageous amount for essentially demographic inventory, with 2 million page views/month they have a $20k/mo business. That does not get them there.

I do have to say that the site is headed in the right direction philosophically, it seems. Here is a product page that actually gives you data above the fold:

So that is the right kind of thing to help them improve frequency. A single 728 across the top and then the content you were looking for. Me, I would move the links to photo galleries at the top to a more compressed area, shrink the logo a bit, and jam the 728 into the very top. Page layout seems really clunky with the 728 where it is.  They picture the “hiver” more prominently than the clothes you came to look at.  Is that bizarre?

Unfortunately, the other direction they are heading to improve frequency is photo galleries, and there seems to be a big focus on driving traffic to the photo galleries on the site now.  No surprise, it is mildly nefarious: Slide shows (their term for photo galleries) are automatically advancing pages that load a 728 in the bottom and a skyscraper on the side. When a slideshow finishes it automatically advances to a new slideshow.  So they can generate a lot of high frequency traffic off a single inattentive visitor.  Of course, this is all worthless traffic. I see a bunch of adify tags, so they are likely seeing < $1 cpms.

I saw back in December that they made an acquisition and in the article, PaidContent references several competitors. I put together a quick Alexa chart showing their traffic ranks:

I remember when Stylehive and ThisNext were the Web 2.0 shopping sites.  It must be Web 2.1 because they are getting pwned by the new kids on the block.  Stylefeeder has moved super quick.  What are they doing?  Couple of simple thoughts, although you never know without testing:

  • Picture first, then description and tags.  Stylehive does tags, then description, then picture, frequently pushing the picture below the fold.  People are shopping visually.
  • Celebs: Stylefeeder gets recommendations from celebrities.  Stylehive from fashionistas.  Celebs probably have a broader appeal.
  • No ads above the fold.  They have a 300×250 on the side and a 728 at the bottom, both below the fold, the result is a cleaner, less cluttered look.
  • Recommendation that shows you other related things that people liked.  Easy to do, why haven’t they done it

Seems unlikely that Stylehive will have the runway to turn this around.  I suspect it is hard to raise more money with such anemic traffic growth in the past year.  Even if they doubled it again next year, they don’t have the revenue to build a real business.  Would Glam buy this business?  Hard to imagine given that they are getting all of Stylefeeders ad space without having to take the risk of traffic generation.

Cogmap API Update July 2008

Tuesday, July 22nd, 2008

Cogmap is pleased to introduce the first significant changes to the Cogmap APIs since their introduction.

These changes incorporate our new URL schemes and also support application authentication.

Here is the low-down:

Organization Chart Data is accessible as an XOXO document by calling a URL formatted as<CHART NAME>

This returns a nested set of unordered lists that describes the organization. Each listed item also contains several pieces of metadata associated with the list. The format is:

<li><a href=”profile.php?id=<personID>”> <FirstName> <LastName> </a> <dl> <dt>Title</dt> <dd><TitleValue></dd> <dt>CogID</dt> <dd><personID></dd> </dl>

  • personID: the ID in Cogmap’s database
  • FirstName: Person’s first name
  • LastName: Person’s last name
  • TitleValue: Person’s title
  • The title of the document is the chart’s name.

So here is an awesome example: XOXO No Inc Chart!

There are a number of other interesting things you can do:

  • If you weren’t sure, for some reason, how Cogmap has changed the name of a company into the URL schema that Cogmap uses, you can tell it to guess by setting the interpret flag.  This might look like: inc?interpret=1 and it will take a shot.  If it fails, or if, in general, Cogmap doesn’t recognize what you are looking for, a 404 is returned.
  • All of Cogmap’s APIs, and for that matter, all of the pages that are protected by access control, can now be accessed programmatically.  On every member’s user settings page there is an “API Key”.  Using your user name and API key will allow you to access protected pages programmatically by adding the following to your “GET” query string:
    • user=<username>
    • apikey=<api key>

All of these APIs still work:

hCards & vCards

  • Every profile page is actually an hCard. Munch it, crunch it, consume it.
  • You can download vCards easy too.<ID of Profile>

Newsfeed API

  • Consume your newsfeed!<your screen name> returns an XOXO XML document of your newsfeed. This makes it easy to turn your newsfeed into something consumed by other applications. Here is my newsfeed as an example.

Revenue Science Proves the Challenge with Lots of Capital

Monday, July 21st, 2008

Valleywag reports that ValueClick is considering buying Revenue Science.  I have blogged extensively about the challenges in raising a lot of capital and exiting appropriately.  This would be a great example.  As Valleywag notes, Revenue Science has raised more than $70 million dollars.  If investors owned 90% of the company, the business would have to sell for $770 million dollars to achieve 10x returns.  If investors own less, the company would need to sell for far more in order to achieve those kind of returns.

What are the odds that the company sells for a price like that?  Well, today VLCK is trading for $940 million.  So even if they sold the business for $250m, they would be getting 20% of a publicly traded company.  That strikes me as unlikely.  A $250m exit, if the investors owned 100% of the business is less than 4x returns.  On an absolute basis, not too bad, but it strikes me as aggressive.  What if they exit for $100m?  They are getting 10% of a billion dollar publicly-traded business.  Pretty good exit!  Probably the investors take basically all of the money and feel like they barely got out.

Bad outcome for all concerned.

Contrast that with Tacoda, raising ~$30m and exiting for $270m.  Huge victory for all concerned by being more efficient with capital and selling out to someone much bigger than the $1b VLCK.

Reviewing Business Plans Before You Show Anyone

Thursday, July 17th, 2008

I talk to entrepreneurs all the time. Frequently, I look at their business plans and advise them on changes that could make them more fundable or likely to be successful. I wanted to document some of those ideas in a blog post that I could refer entrepreneurs to read. This could save everyone a lot of time!

This post is about building financial models for business plans:

I usually build the financial model after I have done the research to write a business plan and after writing a very rough first draft of a PowerPoint describing my business, but before I actually put very much work into the business plan. It drills down to monthly revenue and expenses and has some rollups to quarterly and annual figures. Without a financial model, I won’t have a sense of the kind of sales organization, pricing, or customer volume that I will need to have to be a successful business. I went to Wharton, so I like to look at the numbers.

My financial statements tend to have a similar structure with at least these tabs:

  • Balance Sheet
  • Cash Flow Statement
  • Profit & Loss
  • Profit & Loss Quarterly
  • Profit & Loss Annually
  • Revenue Model (If it is too complex to easily put in the P&L, which it usually is)
  • Headcount – Line by line every employee type, quantity each month and how much they will make each month – this gets rolled up into some headcount and expense numbers used in other tabs
  • Assets – If there are more than a couple of depreciable kinds of things, I will model that up because it doesn’t take me that long.  This is probably not necessary if you just make reasonable assumptions in the balance sheet and don’t have a business with a lot of capex.

Typically, the kinds of business I am interested in have very low capital expenditures – they are mostly about people and software – so my balance sheet and cash flow statements are pretty simple. Very little depreciation, simple AR and AP models, but the basics are there to understand how it has to grow to get profitable, financing requirements, and expenses. I suspect that if you had a business with an extremely complex capital structure, modeling the financials would be even more important.

I typically run my financial model out until I have at least a year or two of profits. Typically, I imagine that I am driving the financials towards some end-state model that represents how we would manage the business once it reached some more mature growth point. This is typically represented, in models I have built, by a revenue growth rate less than 60% year over year and margins around 20-30%. Usually the model goes out four or five years.  If your margins aren’t that high yet, then you probably haven’t reached a state as a business where you can command a strong valuation based on financials rather than perceived upside.  If you are still growing at greater than 60%  and your margins are that high, then you are throwing off so much cash, it always feels unrealistic to me.  When you don’t have to layout as much cash as growth slows, your margins probably vault to some number that seems big to me.

Study your financial model closely. I see the same mistakes time and time again:

  • A large infusion of capital does not allow you to hire rapidly. You will not be able to hire people that quickly. Don’t assume you can.
  • Look for big changes in expenses month over month and ask yourself, “Will I really be able to spend that much more money efficiently thirty days later?” The answer is rarely yes.
  • Don’t assume you can raise capital quickly. Sometimes you can, but this is rare. Assume at least 9 months between financing rounds.
  • Don’t assume salespeople will close deals for the first 90 days they are on board.
  • Don’t assume you will be able to sell all of your web site impressions to advertisers. You need to keep a buffer just in case you have a slow day but you need to fill guarantees.
  • Look at the quarter to quarter revenue growth: Is that realistic? Remember, your job may one day depend upon achieving that growth. Would you bet on it? It may behoove you to slow the hockey stick down a bit.
  • Does it lack a hockey stick? Without an inflection point, it is not really an investable deal.
  • What are your margins in the last year or two? If they are over 50%, then you are probably being too aggressive on revenue or too aggressive on expense management. Why? What good does it do to promise upfront that you will run the leanest, most capital efficient business in the world? Remember, these are commitments you are making to investors. Why not be a little more conservative on expense management (assume you spend more) and a little more conservative on revenue (assume you generate less)? You can still have a bang up business at 30% margins.
  • Are you paying yourself enough? I don’t expect you to have a big salary the first year or two, but when your business gets big and has hundreds of employees, you should be able to go out to eat once or twice a year.
  • Are you paying your CFO and head of sales enough? If you are building a big business, then these people have to be awesome. And if they are, they won’t be happy making peanuts as the business takes off.

It is always the same message: Your expenses are too conservative, your revenue is too aggressive and your growth and expenses are all too lumpy.

Group M CEO Misses the Boat on Google Monopoly

Wednesday, July 16th, 2008

I am sorry, I can’t stop going after posts I disagree with on the web.

The head of Group M complains to Ad Age that if Yahoo isn’t competing with Google, search prices will rise:

“The party who bids highest and who achieves the highest quality score, comprised of price, relevance and likelihood to click, wins. And in a competitive market the price is capped by the incremental cost of the click to the advertiser in search engine A vs. search engines B, C or D and the total volume of clicks that the advertiser wants, needs or can afford.”

There are real issues in pricing in this specific monopoly (read here), but this economics theory argument misses the reality of the situation.  As Aaron Wall already pointed out, breaking down the data, you already pay more for clicks on Google, but that is only because the backend performs so much better.

Rob misses the target further:

“Inevitably, the per-click price of search will continue to rise if other channels deliver less volume and efficiency, and, if not capped by internal competition in the market, they will rise to a fraction below the costs of non-search channels.”

Err, I think the real economic theory here is that the price rises until the market CPA is exactly the same across channels.  For ease of discussion, a simple market like mortgage leads can be looked at.  Let’s say that a mortgage lead is worth $100.  If 10% of clicks turn into mortgage leads, then a mortgage lead firm is willing to pay up to $10/click.  If Yahoo offers worse converting clicks (as Aaron illustrates), then maybe the conversion rate is only 5%, lowering what an advertiser is willing to pay to $5.  If we are theorizing an efficient market, and that is certainly what he implies, then the backend performance will be equal across search and non-search channels.  If it is unequal, then there is an arbitrage opportunity and the advertiser is inefficiently spending in one medium or the other.

Now, search performs great, don’t get me wrong.  People are actively seeking things related to a product!  That is going to perform great.  But the result is higher per click prices in an efficient world, driving to the same back-end result.  This happens regardless of how many search providers there are.

How does internal competition in the market keep per click prices low?  Really, it isn’t about having lots of players, it is about optimization.  Better optimization provides better targeting.  More players doesn’t lower prices.  In fact, by distributing budgets across many algorithms, testing costs rise, creating potential inefficiencies.

Pubmatic AdPrice Index Fails

Tuesday, July 15th, 2008

VentureBeat, along with everybody else, covers a press release about the Pubmatic AdPriceIndex that I love, because it is data-driven, and will now proceed to pick apart because there is not enough data. In summary, this high level analysis of the data doesn’t really tell us enough about what is going on. Pubmatic should let us at the raw data so people can do a more nuanced analysis. Far be it from me to criticize people much, much, much smarter than me (PhDs!), but I think there are probably some aspects of the industry/Pubmatic offering that might not be understood and prevent the data from being controlled in a way that allows for effective analysis of the data set.

Anyway, I know I said in a previous post that I would try to be less harsh to people, but let’s dive in.

The most obvious question is with regard to the most prominent aspect of the press release. They announce that prices went down by a penny month over month. Of course, the first drilled down detail they offer is that small sites (under 1 million impressions) went down $0.32. They don’t provide compositional data, but it would stand to reason, given a data point like this, that small sites may have made up all of the penny shortfall. In fact, they do say that medium and large sites saw their CPM rise.

In fact, as their report spells out, big sites and medium sites have made higher CPMs every month! Maybe the recession is confined to small businesses? Not a chance.

So if we are saying that the penny shortfall was caused entirely by small web sites, the next question is how can we break down those sites. We have almost no insight into this, but one thing we can intuit is that they are working with many more small web sites this month than they did the previous month. If you read the June report, then you see that they worked with 3,500 web sites in that report, whereas they worked with 4,000 for the July report.

I bet most of those are small sites.

So they added a ton of new publishers to their service and prices fell. Web sites that weren’t being monetized previously or were being monetized poorly (either way, probably signs that it isn’t the most valuable inventory) signed up for Pubmatic, dragging price points down. That is not an unfair hypothesis.

One thing you could do is look at how publishers that participated in June saw their prices change month over month.

Alternately, the study indicates that as sites give Pubmatic more inventory, CPMs decline. Maybe they are signing up larger small sites (more profitable business for Pubmatic) and they are yielding lower new CPMs. Some analysis of how CPMs are linked to volume in Pubmatic’s system could explain this.

Considering even more alternatives, every ad network (and Pubmatic’s service) have testing costs. Maybe with the rapid influx of publishers and Pubmatic’s distribution of that publisher inventory over many advertising networks, the yield is artificially lowered by the need to test many new publishers across many new networks. Once again, controlling for the introduction of new publishers is important.

I am tired of typing now, but you can imagine the problems that might lurk in the other parts of the data. Incidentally, I bet that Pubmatic prints these great bio’s of the statisticians to imply that this is good, but I bet that at the very least the Chicago guy is embarrassed to be repping this data:

  • Albert Madansky, Ph.D. is the H.G.B. Alexander Professor Emeritus of Business Administration at the University of Chicago Graduate School of Business, and was the recipient of the 2005 American Statistical Association Founders Award.
  • Michele Madansky, Ph.D. is a media and market research consultant and former VP of Global Market Research for Yahoo!

Obviously, I work at an ad network, so don’t construe this as validation or not validation of what they are saying at all (prices rising, prices falling, don’t know and if I did I wouldn’t tell you).  It is neither.  This is simply a constructive critique of the information they reveal.

Rolling Up Undifferentiated Ad Networks Doesn’t Work

Sunday, July 13th, 2008

Clickety Clack comments on a ZDNet post that there might be a roll-up opportunity among smaller advertising networks. Implicit in his post is the concept that rolling together a bunch of 30-40% reach vehicles might give you some sort of 70% reach vehicle. That might happen, and I am certainly painting in fairly broad strokes here, but a lot of the inventory on these sites is Right Media and MySpace. To my way of thinking, almost anyone can rapidly build a network with 30%-ish reach by doing low frequency buys on MySpace and Right Media.

The real question is what the overlap in their advertising base is. If they actually have differing demand and are not all working the same CPA deals, there could be some leverage to a roll-up there. Unfortunately, there is basically no way of knowing from the outside what that opportunity may be.