SpiralFrog's turmoil, in missives

Below are several e-mail exchanges that were obtained by CNET News during a review of SpiralFrog's rapid rise and fall. For the complete series about SpiralFrog's collapse, please go here and here.

The first chain focuses on Amir Khan, one of SpiralFrog's financial backers, and his skepticism about SpiralFrog's marketing strategy in the summer of 2008. The second thread is a debate between Vesa Suomalainen, SpiralFrog's former chief technology officer, and Joe Mohen, the company's founder and chairman that occurred last fall.

It should be noted that the e-mail exchanges were originally forwarded to multiple people, and copies were subsequently forwarded to CNET from multiple sources. E-mail addresses, phone numbers, and other personal information were removed. Grammar was not corrected. The e-mails were placed in chronological order to make them easier to read.

Editors' note: In the first exchange, the recipients were members of SpiralFrog's board. Orville Hagler, a SpiralFrog employee, was CC'd.

From: Mel Schrieberg
Date: Sun, 13 Jul 2008 09:53:05 -0700
To: Hughes, Al; Amir Khan; Gordon, Bob; Preston, Frances; Gold, Jerry; Joe Mohen; Levin, Jordan; Hutchens, Mark; Strama, Mark; Stagg, Scott; Norcia, Steve; Mackell, Thomas Mackell.
CC: Orville Hagler
Subject: Traffic Analysis

The board committee on traffic analysis that includes Scott Stagg, Amir Kahn, Jerry Gold, Steve Norcia and me, met on the afternoon of Wednesday, July 9th. It was a very productive meeting given the complexity of the topic. As a result of our discussion, the consensus was that we serve two constituencies, that being strategic alliances and the advertising community, and our hypothesis is that these two entities have different criteria on how they view the company. To gain further insight with respect to strategic alliances, we developed a series of questions that I will pose to Jimmy Barge to better clarify Viacom's position. At this point, we believe that we need to grow our monthly uniques while migrating to quality visits.

From a tactical perspective, last Thursday, July 10th, the company achieved its highest daily uniques at 295,228, and we are pacing at 7,615,277. Our current spend is the same as last month, thus we are proportionately gaining traffic virally. I plan to decrease our spend towards the latter part of the month, which will result in approximately 7.1 million monthly uniques. Concurrent to this, I will start migrating to our Phase II quality traffic programs.

Thank you,
Mel

----- Original Message -----

From: Amir Khan
Sent: Sunday, July 13, 2008 4:20 PM
To: Mel Schrieberg
Cc: Orville Hagler
Subject: Re: Traffic Analysis

I am having 2nd thoughts. My friend came to visit the company on friday and he asked very pointed questions to which we had no answers about refresh rate and pages per visit compared to facebook etc., which we keep comparing ourselves to. If u look deeper our traffic numbers are a house of cards. I know him for 20+ years and there is no way he will invest in the company with such metrics. We need to discuss this again.

Amir Khan
Portfolio Manager
3V Capital Management

----- Original Message -----

From: Mel Schrieberg
Date: Sun, 13 Jul 2008 20:15:52 -0700
To: Amir Khan; Hughes, Al; Gordon, Bob; Preston, Frances; Gold, Jerry; Joe Mohen; Levin, Jordan; Hutchens, Mark; Strama, Mark; Stagg, Scott; Norcia, Steve; Mackell, Thomas J.
CC: Orville Hagler
Subject: Traffic Analysis

Amir

To your point we do know what our page views and refresh rates are but you are right we do not know how they compare to Facebook etc, - we are not comparing ourselves to social network sites in that they are not the model we are following - our site does not have any objectable content and therefore we are able to attract the Tier I advertisers where your social network sites can only obtain a CPM of $1;50 and we have averaged $14-16 and recently have an ad campaign with AT&T - their "blue room promotion" where we obtained a $26.00 CPM.

The reason we have compared ourselves to Facebook is strictly one of gross traffic and I strongly believe if we were to make ourselves more of a social network site our traffic would be larger but it would defeat our economic model

I do not think our traffic is a house of cards - our phase one approach was to build traffic as quickly as we could, so that we could attract strategic partners and tier one advertisers and then migrate this traffic which we have stated to do, to enhanced quality programs. There are many sites Rhapsody being one that spends considerably more money on traffic enhanced programs and don't come close to our results. To expand upon this I have spent a considerable amount of time with AOL and Advertising .com executives, (recently a two day meeting in Baltimore with 14 of their executives in attendance) many of which have spent ten years on growing website traffic and they have not seen a site grow so fast. - the key to this is we have been able to converge the discovery and acquisition of music on our site - our pre -qualitative market research determined that this would be critical to our success. Yes, I would like to see greater stickiness to the site but please remember that up until now we have had only one music major. We have conducted three post launch focus groups and the main reason they have not come back to our site was not having a greater selection of music - this is now changing with the ingestion of EMI.

Amir, I would be happy to met with your friends again and clear up any miscommunications, I found them to be very bright and articulate and we might have not made ourselves clear on our business model; Your point is well taken on better site metrics, we are working on improving this area but as I stated at out committee meeting we can only determine the number of clicks for each of our programs (receive these from the vendors - Google, Yahoo, MSN, Advertising ,com etc.) but we cannot measure these as accurately as would like to with respect to the method Nielsen measure uniques - would take additional funding to outsource this capability

Thank you

Mel

----- Original Message -----

From: Scott Stagg
Sent: Sunday, July 13, 2008 11:54 PM
To: Mel Schrieberg; Amir Khan; Hughes, Al; Gordon, Bob; Preston, Frances; Gold, Jerry; Joe Mohen; Levin, Jordan; Hutchens, Mark; Strama, Mark; Scott Stagg; Norcia, Steve; Mackell, Thomas J.
Cc: Orville Hagler
Subject: Re: Traffic Analysis

I agree with mel and disagree with amir. We all are in agreement that our first objective was to grow uniques which we have done and there is clearly some disagreement about what our next step is, whether to continue growing uniques which the advitisers seem to want, or slow the unique growth and increase quality. To me, it is all about impressions and if the choice is to get there with more uniques and less pages turned vs less uniques and more pages turned, with both equaling the same amount of impressions, then I would choose the higher uniques.

Mel, could you please try to find out how many pages visitors average on facebook to potentially answer amirs friends concerns.

And amir, could you tell us exactly what you and your friends concerns are and what info you need from mel. Many people have put in an extraordinary effort to get where we are and this is far from a house of cards. Sent via BlackBerry by AT&T

----- Original Message -----

From: Amir Khan
Sent: Monday, July 14, 2008 2:19 AM
To: Scott Stagg; Mel Schrieberg; Hughes, Al; Gordon, Bob; Preston, Frances; Gold, Jerry; Joe Mohen; Levin, Jordan; Hutchens, Mark; Strama, Mark; Norcia, Steve; Mackell, Thomas J.
Cc: Orville Hagler
Subject: RE: Traffic Analysis

Mel & Scott,

My concerns stem from the following:

1. We are embarking on a massive capital raise. The investors we are and will be encountering for this round will be more sophisticated and we are having a hard time convincing people of the valuation of the company even after signing 2 majors and all the independents. If we are discussing raising money using the same instrument as our round last may, then clearly something is wrong.

2. Our entire financial model is based on a steady state mid teens pages turned per visit out in the future. While we are managing our uniques per month number pretty well, we are currently running at 2.6-2.9 pages per visit (public info as per alexa) instead of our presentation predicting 6-7 in July to Dec phase, with correspondingly low time spent on site. Therefore our available impressions are running at 40-50% of the model we are showing to investors. Just as we seem to be targeting uniques, we need to manage the pages per visit as well otherwise it will call into question the entire financial model. If we keep growing uniques per month but pages per visit lags, guess what - impressions served will be well below model projections. To put it bluntly, while we are 15x our Oct 07 uniques, we are just 2.5x Oct 07 impressions served. If I were looking at the company afresh, I would not buy the improvement in uniques quality assumption in our model without assuming a fall in uniques. Its just not possible!

3. The people we seem to be attracting to our site from the affiliate marketing programs are NOT interested in music. Hence the low registration rate, pages per visit, time on our site, high bounce rate. I refuse to believe that people in the advertising world and the potential acquirers will not see this as buying traffic. I remember Jerry on the conf call could not digest that either. When my friend, Jassi, who recently sold his company at a nice price, and understands this space stayed focused on the quality of the uniques and we did not answer his question directly, I got more concerned. If we are going to compare ourselves to facebook, youtube in terms of total traffic, we must expect questions regarding the quality of traffic as well. Anecdotally, Saurabh, our intern last year who was a big facebook user turned probably 25-30 if not more pages per visit. I doubt if facebook, myspace, youtube or any of the beacons of the new internet age have ever relied on affiliate marketing. Their VC sponsors would not have allowed it. Finally I believe its entirely possible that youtube/myspace went to acquirers and told them we don't have a salesforce but you do and you can make this much money from us. By having a salesforce with 4-5 senior guys plus a bunch of juniors and not having a decent pipeline, are we telling the world that perhaps the model is faulty? Even if we assume we get a 15 CPM and myspace gets a 1.50 CPM, and if we have 2.6 pages per unique vs 25 per unique for them, we end up with the same gross revenue with the same uniques but we would have a bloated marketing cost, where they would have none. Add to that the fact that their inventory costs them zip, and we have large upfront acquisition costs and we end up keeping just 33% of our revenue (if we are lucky to use up the recoupable advance) hence on a net revenue basis we would be 1/3rd their net revenue with a bloated marketing expense. So before we start dreaming of youtube valuations, we have to get our perspective straightened out.

4. The investor presentation calls for a 2008 marketing spend of 7.9mm. If last month our total spending was 1.1 or 1.2mm, and we add to that what we spent in the other months so far, we will probably have just 3-4mm left for the rest of the year or 500k-670k per month (and we must include all sorts of marketing in this calc). If we require 1.1-1.2mm to get us from 5mm to 6mm, how come we will require lesser to keep increasing uniques?

5. If we keep buying traffic and we do ramp uniques to 10mm, we will have an even harder time bring up pages per visit from under 3 to plan, as compared to doing it now. We must improve the quality now, and stabilize traffic rather than aim for 7mm+. We can not become uniques addicts at the expense of quality. We got it above 5mm, let's just manage it within a range, all the while replacing affiliate programs for better programs.

6. If spiralfrogclub generated just 10,000 emails, and the Alicia Keys program (from which we expected 4mm new uniques) generated just 3,000 new uniques, we should make sure we don't commit the same mistake again. These viral programs accomplished just 1 thing - they made me sick. I am unwilling to see a single program we have in the works fail. And there have to be accountability and repercussions for any marketing failures going forward. I don't know whether we pulled resources from other projects to work on these as well so besides money; opportunity cost, manpower redeployment etc also need to be accounted for. Did these happen at the expense of release 3.0 or other enhancements which would make new visitors stick to us? And with the affiliate programs, are we doing the same?

7. If google adwords leads to more registrations, these people will probably come back and hence we should look at cost per pages visited by unique per program not just cost per unique. If we did that chances are the 28 cent to 8 cent comparison will look a lot closer. Furthermore don't we think search engine optimization is a better tool to drive people who are actually looking for music than affiliate marketing? I also believe we must do away with the email requirement. We can let the user choose a user name/password and let him decide if he wants to give us his email/tel no for email/text updates etc. That will be a lot more acceptable in the age group we are trying to serve. I remember arguing about this back in Dec with Roger Munford and I still don't think people like to give out email addresses for fear of phishing etc. If we are competing with pirate sites, we must understand the behaviour of our target audience.

8. Whatever happened to alliances with mp3 player manufacturers, cellphone companies, and generally anybody who would benefit by pushing us? Should we not have banners on their sites or make Samsung mention "compatible with spiralfrog.com"?

9. Our first priority has to be content acquisition, at the expense of every single thing including sales, uniques, salaries etc. I also believe we should line up a buyer who would buy us provided we got Sony BMG and WMG. With capital markets being the way they are, we can not rely on the assumption that a 20mm round is definitely going to happen. In my mind, its very hard to justify new hires, more marketing spend as time goes by especially with the markets being the way they are.

Having said all that, its not too late to rectify our message to investors to reflect what we have been doing. Our model needs to be more believable and yet we need to show a better sales pipeline. A 2.4mm pipeline is just not acceptable after having hired so many people in sales lately. Its possible that a strategic out of the 20 possible strategics will value our uniques differently but we cant believe that. We must be able to convince more people - I hv heard too many comments about potential investors peeling back the layers and not being impressed with uniques quality.

AK

Editors' note: The e-mail exchange below occurred last fall between Vesa Suomalainen, SpiralFrog's chief technology officer, Joe Mohen, the company's founder, and Matthew Stern, vice president of marketing. Others who received the e-mail were a combination of managers from SpiralFrog and 3V Capital Management, the start-up's main financial backer.

From: Vesa Suomalainen
Sent: Thursday, September 25, 2008 1:38 AM
To: Jesse Paynter; Matthew Stern
Cc: Joe Mohen; Amir Khan; Michael Puccini; James Campbell
Subject: business case for paid search?

I have not seen this written down anywhere, so I thought I'd take the initiative so we can allagree on the numbers before making a decision whether it makes sense to continue spending money on paid search on Google, Yahoo and MSN adwords. Please feel free to correct if I got any of this wrong.

We've been spending up to $600K on paid search per month over the last several months. For the most recent month that I have detailed numbers (June), we paid roughly $530,000 for 2.65 million.adword clicks, resulting in just over 140,000 user registrations. Each click cost us approximately 20 cents, and the regitration rate from paid search was roughly 5.4%.

This means it is costing usabout $3.70 per registered user obtained through paid search. At an average 6.5 pageviews per visit (latest month stats from Google Analytics), we're earning roughly 3.25 cents per user visit in ad revenue (assumptions: 2 ads per page, average $5 CPM, 50% of ad inventory sold each month). It would take 114 visits per user to earn back the money spent on paid search to break even. And that's assuming we'd keep 100% of the revenue - not figuring the 66% royalty due to the music content owners.

It is therefore difficult to arguevia a business case to continue funding paid search. The per-click rate is simply too high for an ad-supported business. Paid search ads are primarily designed for businesses that sell goods for relatively high prices - using paid search as a method to drive traffic for an ad-supported site would be difficult to support even if the conversion rate (from click to registration) would be 100%.

The obvious concern is what would happen to our traffic and ad sales if we stop paying for adwords? For one, our unique visitor numbers would drop by the corresponding number. However, since nearly 95% of paid search clicks end up in just one or two pageviews, our average pageview count would go up dramatically - and the # of total pageviews would drop but not by the same ratio as unique visitors. There's clearly a cost to this, but it's very difficult to argue that the downside is worth the guarantee of losing money on every dollar spent.

The other statistic that ought to dramatize the paid search cost number is this: 12 months after the launch of the site, where one would expect numbers to have stabilizedfrom the early peak in marketing spend and the initial delay in sales revenue, we're still spending upwards of 300% of our revenue in marketing costs (I don't have the latest sales figures so I could be off here a bit) - whereas e.g. Napster which was losing money every quarter due to their high marketing spending was using about 25% of their revenue to drive traffic.

There seems to have been universal agreement that our affiliate marketing program (Platform A) was not appropriate for our site given the high cost and the low conversion ratio - but I'd suggest the paid search is not much different. It will be difficult to defend this amount of sustained marketing spending to a new incoming investor who's going to scrutinize our business model and numbers carefully before making an informed investment decision. There are other, more inexpensive methods to drive traffic that we should pursue - e.g. traffic sharing programs with other sites or lower-cost pad clicks with music discovery sites. Paid search has an obvious cost incompatibility with our pennies-per-visit banner-ad-based business model.

----- Original Message -----

From: Matthew Stern
To: Vesa Suomalainen; Jesse Paynter
Cc: Joe Mohen; Amir Khan; Michael Puccini; James Campbell
Sent: Thursday, September 25, 2008 2:31 PM
Subject: RE: business case for paid search?

I agree with the sentiment (decrease dependence on paid methods while searching for alliances - barter or otherwise), however the metrics have been getting better since June. I will do a full reconciliation at the end of the month, but judging on pacing so far, this is what we see:

September will see the highest ever monthly membership total (30% higher than June) yet September versus June member acquisition costs are 63% less expensive September will see the highest ever impressions served (16% higher than June) yet September versus June cost per impression are 58% less expensive Page/visit September versus June are 70% higher (7.45 vs. 4.39) Time on site September versus June are 140% higher (8 minutes vs. 3:20)

NOTE: The improvements are from finding alternative and less expensive banner programs that also deliver better metrics versus Platform A's incentivized traffic. Also, I do not include the NFL costs in September - the $200,000 spend netted 8,175 clicks and 407 registrations.

Finally, to bring this full circle, there have been fruitful discussions with like-music sites to swap traffic and share revenue...stay tuned.

----- Original Message -----

From: Vesa Suomalainen
To: Matthew Stern
Cc: Amir Khan; Peter Desloge; Michael Puccini; James Campbell;
Joe Mohen; Jesse Paynter
Sent: Thu Sep 25 18:21:50 2008
Subject: Re: business case for paid search?

I'm afraid you're missing my main point. We must adjust our user aquisition costs and methods to fit our own revenue model for it to make sense. We've now proven that any amount spent on paid search or other cost-per-click advertising is guaranteed to lose us money. The oher point is that the amount spent on marketing in a given month should be a sensible proportion of our own ad revenue - not a multiple of it.

According to a spreadsheet by our SEM vendor Bluemark, we've paid for 15.6M adword clicks across Google, Yahoo and MSN since February 2008 at an average price per click of just under 20 cents. This amounts to a total of $3.12M. This number is astounding - and does not include the other marketing programs (Platform A, Alicia Keys, NFL etc. etc.).

The NFL promo seems to have a set a new record for our user acquisition costs - $491 per registered user. Paid search looks rather attractive in that light....

The other obvious consideration is our present funding situation. When we're having trouble paying emlployee salaries and basic expenses, it is the wrong priority to spend money on marketing programs that can at best be called loss-leaders. In my view, we should hunker down for a while and get our content catalog and site features enhanced, build no-cost traffic sharing partnerships, establish a strategic alliance or two, get more funding and then relaunch.

----- Original Message -----

From: Joe Mohen
To: Vesa Suomalainen; Matthew Stern
Cc: Amir Khan; Peter Desloge; Michael Puccini; James Campbell; Jesse Paynter
Sent: Thursday, September 25, 2008 6:32 PM
Subject: Re: business case for paid search?

The only caveat I would throw in here is that paid search would make sense if the acquired users were to return to the site more often. Put another way, if our 2.4MM registered users came back twice a week, then we might view this differently.

While u r correct with the NFL digitally there were also approx 6 TV spots that went with it, and that mitigates this, although we can't say how much

----- Original Message -----

From: Vesa Suomalainen
Sent: Friday, September 26, 2008 11:48 AM
To: Joe Mohen; Matthew Stern
Cc: Amir Khan; Peter Desloge; Michael Puccini; James Campbell; Jesse Paynter
Subject: Re: business case for paid search?

I believe we have established that we end up losing money even if the users acquired through paid search keep returning to the site.

To be fair, let's adjust my calculation below to grant that registered users end up spending more amount of time on the site than the average - which is skewed downward by those paid search users that end up quickly leaving the site. From Google Analytics stats for October 2007 (when we had no paid search programs ongoing), the average pageviews per visit was 14 and # of visits per user per month was 2. These numbers are very close to our original business plan assumptions.

Let's even use an average of 15 cents per paid search click. And a generous conversion rate of 10% to become registered users. And blindly assume all registered users end up becoming permanent members. Each such user would cost us hence $1.50. At 14 pages/visit we'd earn 7 cents per visit in ad revenue (2 ads per page, $5 CPM average, 50% of ad inventory sold). Next, we'll need to adjust this to subtract 66% royalties due to music owners - leaving us 2.38 cents per visit. It would therefore take 63 visits from each such user for us to break even. At an average 2 visits per month, we'd need each user to become an active SpiralFrog member for over 5 years to make this work financially. Our present stats show that a registered user stays active an average of less than 2 months.

We can therefore conclude that paid search and cost-per-click advertising is permanenty incompattible with our business model. Even with the generous assumptions made above, there's no way to make the numbers work in our favor.

It would be helpful if someone could send a note indicating our ad sales per month and % of ad inventory sold each month. For too long, we've lived in an environment where some of this data was kept secret, disallowing us to make rational decisions elsewhere. I've heard that one of the main arguments in favor of continuing with paid search is our ability to earn ad impressions already sold - but if you believe in this logic, it's akin to spening more money so we could keep losing even more.

----- Original Message -----

From: Joe Mohen
To: Vesa Suomalainen; Matthew Stern
Cc: Amir Khan; Peter Desloge; Michael Puccini; James Campbell; Jesse Paynter
Sent: Friday, September 26, 2008 8:54 AM
Subject: RE: business case for paid search?

I think it is more complex. While I more or less agree with your points, there are more variables.

First, is our ad sales are in their infancy. It is too soon to judge marketing costs relative to them.

Secondly, it is true that you need a certain amount of "uniques" to get the attention of customers and key industry players. If our uniques fall below a certain level, potential customers will not take our meetings, and partners will not take us seriously; even Viacom might walk away. You will not get into meetings. While most of what Mel did was insane, there is some truth to this point.

Thirdly, the key business variables are how often users come back to our site. We need more data points on this. I think it might make sense for you to go through all the FEEDBACK emails (starting with right after EMI was added), to see what people are telling us about our site, what they like about it, what they do not. You did a good job of this late last year, and it would be interesting to see what they are saying now.

It might also be useful to read the SpiralFrog blogposts, to see what real users are writing.

----- Original Message -----

From: Vesa Suomalainen
To: Joe Mohen; Matthew Stern
Cc: Amir Khan; Peter Desloge; Michael Puccini; James Campbell; Jesse Paynter
Sent: Friday, September 26, 2008 10:02 AM
Subject: Re: business case for paid search?

For a while, I guess we all were sold off on the "mometum theory". The belief was that if we demonstrated solid user growth and increased # of uniques, it would open more doors for us at advertisers and music labels and amongst the press and music industry. The cost did not matter since the exposure would be temporary and we would switch from paid to organic growth in a matter of months if not weeks.

I started arguing in late spring that this is, if not outrigth cheating, at least self-deception. We were claiming super unique user growth while we knew we were just getting users to bounce off our site. Our approach was not far from hiring internet users in India to click on our home page to get the uniques # to continue growing. Anyone who'd ask us direct questions about average # of time spent on site or average # of pageviews, or retention of registered users would immediately find out our little secret. And these figures stay permanently in our books for incoming investors to look at and ask us after-the-fact. How do we explain spending $1.5M in marketing in the month of June when our resulting revenue was $69,711? An ooops?

This is our opportunity to start fresh and blame the months passed as a failed experiment. I'd argue that our main problems are lack of music content and lack of site features - fixing these two is key to our future success. If there's one aspect I'd like to emphasize above all is that the unbalanced and incorrect focus on paid marketing is sapping our resources for money, development resources and management attention - besides being counter-productive to our overall goals.

Next week, we hope to add the IODA catalog to our live site - this is 1,081,028 additional tracks, bringing us close to the 3M track marker. And this aggregator catalog is multiple times more relevant than the similar-sized Orchard label which is mostly unknown European music. Do we have a marketing program ready to go when this catalog goes live?

And our main problem in getting these 1M tracks online quicker? Finding a soul in this company who's able and willing to pay for the $2,000 that it cost us to buy the hard drives required to ship the content from Seattle to our production data center in Ashburn. I find it somewhat ironic to be discussing spending priorities of several millions and a few thousands in the same email, but here you go.