I Like Peanut Butter

Recently many of us read with interest Yahoo! exec Brad Garlinghouse's leaked "Peanut Butter" memo on what the venerable portal needed to do to revitalize itself.  In sum, Garlinghouse argues that Yahoo! needs to narrow the set of businesses and initiatives it's pursuing, and clarify/simplify/shrink the organization around the surviving priorities.  "I hate peanut butter.  We all should," says Garlinghouse.

Gutsy note, if awkward leak.  Clearly correlated with, if not fully causative of, the subsequent changes at Yahoo!  I don't work there, so it's hard for me to judge whether the shape and extent of the challenges there specifically warrant the medicine he prescribes.  But I'd think about it differently: peanut butter is good sometimes.

All companies, especially ones like Yahoo!, are a collection of opportunities.  Some are "here-and-nows" and others are "maybe-laters".  You need both, but you can't manage them the same way.  Because windows in web-land are short and uncertain, when you've got a good deal going it makes sense to focus on exploting it fully while ignoring distractions as completely as possible.  At the same time, when the deals you're juggling are iffy, it pays to spread bets, not only to hedge them but more importantly to learn.

Let's look at the competing initiatives Garlinghouse referred to in his memo:

(Note: YME = Yahoo! Music Engine, and YMG = Yahoo! Media Group)

• "YME vs. Musicmatch

• Flickr vs. Photos

• YMG video vs. Search video

• Deli.cio.us vs. myweb

• Messenger and plug-ins vs. Sidebar and widgets

• Social media vs. 360 and Groups

• Front page vs. YMG

• Global strategy from BU'vs. Global strategy from Int'l"

I live in two worlds, the one inside the "digital bubble" and the one outside it.  When you live inside the bubble, and Garlinghouse is closer to its center than I am, you're tempted to think that the classes of applications and use cases these initiatives represent are givens that have "crossed the chasm" already, and that within these classes  we've largely converged on the most usable approaches to implementing them for the people we're targeting.   But when you live outside the bubble, it's not clear that's the case.  In certain circles I travel in, 99% of the people I meet will understand what Del.icio.us and MyWeb are, and the differences between them.  In others, 99% will smile sympathetically and think I'm speaking in tongues.  More importantly, they may represent different segments that might differentially value social vs. tribal tagging solutions, and doing away with peanut butter in this instance might foreclose what might become valuable opportunities to segment the market.  In this case at least, I think it's probably too early to know.

So what?  How should a big media company with businesses all across the maturity spectrum organize itself?  To me a metaphor that makes sense is the idea of shifting a car up through its gears to accelerate when it's moving at different speeds.  It needs low gears to make it possible to start from a standing stop and find its way slowly through local neighborhoods, and it needs bigger gears to keep its engine from exploding when it's taking advantage of pre-existing momentum to move faster on the highway.  Plus, it needs a good clutch to handle the shifts smoothly. 

Enough with the Zen. 

At any given point in time a digital media's business will be supported by one or maybe two big properties or business models, like search and/ or display ads (as is the case for Yahoo!).  These need to run in fourth or fifth gear all the time -- responsible execs, managers, and individual contributors need to focused on them 90% or more of their working hours.  Organizationally a firm may need to be strongly functional for them, to make sure execution squeezes the most profit possible from them. 

Behind these are a collection of second- and third-gear businesses:  niche businesses, with high-six to high-seven-figure revenues, in investment mode or at minimum profitable and helping the organization learn important things, if not otherwise commercially promising.  These run with leaner and less-functionally-separated management structures, allowing them to change direction more quickly to figure their way forward.  It's ok to have multiple ventures here competing in the same concept space, especially if we think we haven't yet made the leap from power users to the general public, or if we foresee multiple segments with different needs to be exploited.  In this latter case it's especially important to have multiple ventures pursuing multiple segments, since it's unlikely a single small-scale business can do that as effectively.  As for how many make sense, it's instructive to look to the average venture capital firm, where the odds of success and the degree of involvement from the VC's drives portfolios numbering on average 27 companies covered by on average 6 partners, or 4.5 per partner.  For a strategic investor that may want more control/ less syndication, that number might be more like 3 per "partner" (stats welcome here if any reader can point me to some).

Finally, there's first-gear opportunities:  concept-and prototype-phase ideas, with the bright line between the ones to ignore and the ones to watch being users.  Here, a Google-like "20%" program, supplemented with a little cash to pay for the server bill and a few AdWords to help drive traffic and go to relevant conferences, makes the most sense to me.

A program like this carries with it some ancillary organizational requirements.  You need to recruit, manage, and reward differently in each stage.  You have to be careful not to exalt the culture of one stage over another.  Approaches to and integration of different  disciplines -- tech, content, business, design, user management -- vary significantly across the stages.

Since it's hard enough to get these things right at any one stage, an alternative approach is to decide to specialize in a particular stage.  Explicitly or implicitly, this is how the natural ecosystem seems to shake out today: startups, VC-backed independents, public "strategics".  Firms frequently get into trouble when they ignore the boundaries around their chosen stage(s).  The key here I think, especially in a quickly-evolving landscape like the one we now live in, is for firms to acknowledge chosen stages explicitly, tune their "boundary" ops for generating and "promoting" opportunities as carefully as they tune their "core" ops, and cultivate great networks of partners pre- and post- their chosen stages.  In some cases these partners might be separate divisions of the same company; in others, they might be separate companies entirely, or jv-hybrids if a firm is good at managing such models.

How would I have re-organized Yahoo! then?  I don't know.  But I can say I would have left room for some peanut butter at breakfast and lunch.