Google and Dell recently announced a major distribution deal. Reportedly, Dell will install on all its Business and Home PCs the Google Toolbar and other applications, in addition to making the default homepage a co-branded Dell-Google homepage. Unfortunately, very few other details have been disclosed.
In an investor call last week, GOOG management stated that the deal contained no hardware barter component. And management asserted that the 6-month trial generated both increased search usage by existing Google users and an increase in Google search users. Indeed, GOOG executives went so far as to say they were “ecstatic” with the results.
Our channel checks indicate that there are two revenue streams to the deal – a per activation fee and a search advertising revenue share. But we have not been able to confirm any other specific details. Given the importance of this deal, we thought it key to develop an independent economic analysis of it. But with so many unknowns – the most important of which, in our view, are the exact per activation fee, the number of Dell PCs likely to be activated, the extent of the incremental searches generated over the 6-month trial, and the revenue share agreed to on the resulting search revenue – we decided to do some reverseengineering.
Instead of using a range of assumptions to try to spot-predict incremental revenue, EBITDA, and EPS to Google, we reversed the process and tried to determine how many incremental searches would the deal have to generate in order for GOOG to:
1. Recover reasonable activation expenses; and
2. Produce enough revenue to make the deal neutral to Google’s current 64% EBITDA margin structure.
We believe the answers to these 2 questions are what are most important to GOOG investors. Our full detailed analysis is available to clients, but to cut through to the main elements of the sensitivity analysis, we focused on what we view as the two key levers to the deal’s economics for Google – what traffic acquisition cost [TAC] or revenue share will Google be providing to Dell and how much will Google pay per each Google software package activated on new Dell PCs.
That means, we held several variables constant. First, we assumed a consistent 13% click-thru rate – in-line with what comScore currently estimates for Google. (Click-thru rates have been rising, so our consistent 13% assumptions layers in extra conservatism.) Second, we assumed a $0.55 price-per-click – in-line with our current industry assumptions for 2006-2008. (We assume $0.50 for 2006 with a 10% increase each of the next two years.) And third, we assumed a 70% activation rate on the Dell Home PCs and a 20% activation rate on the Dell Business PC’s Citigroup’s Hardware Analyst Richard Gardner is projecting for the next three years. (And these activation rates were gut-checked with Richard.) All in, this maths out to 49MM activated PCs over the 2006-2008 timeframe.
What’s the gut-check on the 60%-80% revenue share that we use in our sensitivity analysis? Google’s current TAC is 78%, and we assume a lower TAC as part of this deal given the upfront activation economics. What’s the gut-check on the $2.00-$6.00 per-PC activation fee?
Reports that prior distribution deals with HP (HPQ) were in the $2 ballpark and Google management comments that this deal was more extensive than prior deals. We present below our specific calculation of Dell deal economics for Google under a breakeven scenario [click to enlarge].
So now for the results:
1. Recover reasonable activation expenses: Our sensitivity analysis below indicates that in order to cover reasonable activation fees over the next three years, the deal would have to generate between 3.4B and 20.4B incremental searches. And our best-guess mid-point is 9.1B incremental searches, assuming a 70% TAC and a $4.00 per-PC activation fee. How reasonable are 9.1B incremental searches for Google? In our view, very reasonable, given that we estimate Google generated approximately 90B searches on its partners’Websites in 2005, and by our estimates is likely to generate 335B searches on its partners’Websites over the 2006-2008 timeframe. So all in, that 9.1B bogey implies less than 3% incremental search volume for Google. With Dell being one of the leading PC manufacturers, with Dell being a top 10 U.S.Web retail property, and with GOOG management being “ecstatic” over the 6-month trial, that 3% seems very reasonable.
Incremental Deal Searches [B] Required to Cover Activation Expenses
To fill out the numbers, we calculate that our mid-point guestimate of 9.1B incremental searches would generate $194MM in incremental net revenue to Google over the 2006-2008 period, but no impact to the bottom line given that this would solely cover activation fees.
2. Produce enough revenue to make the deal neutral to Google’s current 64%EBITDA margin structure: Our sensitivity analysis below indicates that in order to produce enough revenue to make the deal neutral to Google’s current 64% EBITDA margin structure, the deal would have to generate between 22.1B and 57.9B incremental searches. And our best-guess mid-point is 34.1B incremental searches, assuming a 70% TAC and a $4.00 per-PC activation fee. How reasonable are 34.1B incremental searches for Google? In our view, stretchy but possible.
Again, we estimate that Google generated approximately 90B searches on its partners’ Websites in 2005, and by our estimates is likely to generate 335B searches on its partners’ Websites over the 2006-2008 timeframe. So all in, that 34.1B bogey implies slightly over 10% incremental search volume for Google. With Dell being one of the leading PC manufacturers, with Dell being a top 10 U.S.Web retail property, and with GOOG management being “ecstatic” over the 6-month trial, that 10% seems possible, although we would view it as stretchy or relatively challenging.
Incremental Deal Searches [B] Required for GOOG to Maintain 64% EBITDA Margin
To fill out the numbers, we calculate that our mid-point guestimate of 34.1B incremental searches would generate $540MM in incremental net revenue to Google over the 2006-2008 period. Against our assumed $194MM in activation fee payments to Dell, that would generate approximately $345MM in incremental EBITDA or approximately $0.70 in incremental non-GAAP EPS over the 2006-2008 period.
Quick Overall Takeaway: When reports first surfaced earlier this year of a potentially large Google Dell distribution deal, it was partially perceived as a significant negative development for Google – a step up in its operating expenses and a defensive move against pending competition from Microsoft in the form of the integration of search intoWindows Vista and the tighter integration of search into Microsoft’s new browser, IE7. While we would still view the deal as a defensive move, we believe the sensitivity analyses above imply economics to Google that are more favorable/less negative for Google than the market realizes.
Specifically, we believe that the 3% incremental searches required – using mid-point assumptions – for Google to achieve break-even on this deal appear very reasonable. And we believe that the 10% incremental searches required – again using mid-point assumptions – for Google to maintain its current very high margin structure are stretchy, but possible. Very likely, we believe, the Dell deal will generate material new revenue for Google – in the $400MM-$500MM range over three years, with modest EBITDAmargin dilution and approximately $0.40-$0.60 in incremental non-GAAP EPS over the 2006-2008 period. That would seem to be a decent deal for Google. And probably explains why Google is eager to do more of these.