Archive for the open source Category

A Scalable Approach to an Unwieldly Medium 

Since leaving my former position as Director of Mobile at OMD to take on a new role inside Microsoft’s Mobile Advertising group a little over a week ago, many of my agency colleagues have asked for my advice on how mobile can become a center of both excellence and profitability for their respecitve firms.   While I’m always happy to dish out ad hoc learnings to anyone with a friendly ear and a pint of Guinness, I thought it would also be valuable for me to share my formal approach to this specific challenge at OMD.   

Bringing OMD into a leadership position in the mobile media landscape would be the most formidable challenge of my career, but before I get to it a little backstory is needed.   Immediately prior to taking on the OMD gig I had a few weeks of downtime by way of a two week honeymoon cruise in the Baltics.  In between the cocktails, Copenhagen and the unmentionables I was also squeezing in a little dork time in the form of some suggested essay readings, mainly focused around the open source software movement.   I was immediately struck with the similarities between a large scale software development process and of my own upcoming mission at OMD… and soon there after the approach and resulting action plan below began to take shape.


“Plans are useless but planning is indispensable.”  – Dwight D. Eisenhower

Like any good plan, mine was a living document.  But while some tactical details may have become casualties of my shifting priorities, the plan’s core principles, objectives, strategies and overall approach remained remarkably true throughout the process.  It was these primary processes that made up the heart of my approach to the challenge at hand, and it is this that I mainly wish to share.

The document below contains most of the components of my plan, in the form of a “Statement of Purpose” to senior agency management.   While some called it my “Jerry Maguire” moment, and others had fun needling me with “Cathedral”, “Bazaar” and “Sisyphus” references for the following year, I think most will agree that it led to a highly productive period for both the agency and its clients with respect to mobile media.  In the following months I also learned that the approach is also not altogether unique, as at least one other media agency (Razorfish) utilizes a similar distributed approach to emerging media, including mobile.  That said, I firmly believe in the approach outlined below, and encourage others to experiment with the strategy if faced with similar challenges.

Enjoy… and of course I welcome your feedback!




The Mobile Advertising Bazaar

A Media Agency Blueprint for Monetizing Mobility



While there is widespread agreement that rapid prominence of the mobile device around the globe represents a tremendous long term revenue growth opportunity for digital advertising agencies, to date the vast potential of the channel has gone relatively untapped.   While excuses and marketplace misperceptions abound, simply put the inability of our industry to aptly monetize the mobile channel can be best traced to a fundamental failure of organizational philosophy.

This is not to say that agency monetization of mobile channel is not fraught with significant barriers to success -far from it.  Anyone with even a rudimentary knowledge of the channel can point to any number of seemingly insurmountable challenges in the mobile advertising and marketing sectors:  Marketplace fragmentation, lack of standardization, poor infrastructure and limited reach are among the most prominent.  Yet it is preciously these challenges that necessitate a vastly different approach to mobile than the flawed models employed by our competitors. 

The following pages provide just such an approach, a recommendation set that isolates and circumvents the trappings of previous systems by drawing upon the successes and learnings of industries faced with similar challenges, the sum of which form an agency model for the successful monetization of the mobile channel that scales on a national, regional and global level.    


Key Challenges

While the barriers for the media agency wishing to successfully monetize the mobile channel are too numerous to quantify in the limited confines of this work, chief among them are the following key challenges – all of which must be overcome if the agency is to activate the channel on behalf of its clients.

  1. Marketplace Fragmentation.  The mobile advertising and marketing landscape is besieged by an expanding variety of tactical channels with few dominant players and frequent new entrants.  The sheer volume of mobile advertising publishers, networks, infrastructure providers and other entities wishing to capitalize on the ever evolving sphere of mobile opportunities bog down planning cycles and overwhelm the creative process.
  2. Lack of Standardization.  An intimidating array of violently competing technical and advertising formats fueled by powerful marketplace gatekeepers and evolving business models confound the strategic planning process and hamper the swift planning and execution of mobile advertising initiatives.
  3. Poor Infrastructure.  The mobile channel is further distressed by comparatively limited research sources, rudimentary planning, ad serving, tracking, optimization tools, and an inherent inoperability between mobile publishers, networks, operators and key channels – all  of which often necessitate redundant and/or manual operating procedures, ultimately adding incremental labor stresses to the entire system.
  4. Limited Effective Reach / Low Perceived Value.  While varying significantly by tactic and geographic region, overall there is a general marketplace consensus that the reach potential and creative options of mobile marketing and advertising channels lag considerably behind most forms of established offline and digital media. 

Certainly many of the aforementioned issues are inherent in any emerging medium, however one could make a compelling argument that (with respect to mobile) the intensity these challenges has escalated to a scale that dwarfs previous “like” emerging media scenarios; Whether driven by the sheer technical complexity of the channel, global ubiquity or simply soaring investor interest, the frenzy surrounding the emergence of mobile as a marketing medium has inflated these barriers by orders of magnitude beyond what has been previously experienced by most other forms of “new” media. 


The Cathedral and the Bazaar

In searching for a model that would satisfy these vast hurdles, inspiration has been drawn from a seemingly dissimilar industry that has faced a very similar set of challenges: The software industry. 

As with the mobile marketing and advertising system, the software development process has been similarly stymied by a confluence of like phenomena: competing technical standards, dizzying market fragmentation and inadequate toolsets.  Like the marketer, the software engineer also struggles with a complex problem-solving process straddling both creativity and analytics, rooted in real time business objectives, aggressive deadlines, and overarching cost constraints – all driven by multiple objectives and taskmasters. 

While challenging on any level, these issues become increasingly prominent when attempting to scale the software development process to meet the accelerating complexity of global software needs.   This challenge was most notably documented by Fredrick Brooks in his 1975 breakthrough tome “The Mythical Man Month,” where the author famously identifies the core paradox that has become known as “Brooks’s Law.”  Put simply, it follows that in software development, like most “tasks with complex interrelationships… adding more men [to a project]… lengthens, not shortens, the schedule.” 

This apparent contradiction is due to escalating internal communication needs and the inability to fully partition the labor demands, as related to the many complex and interrelated tasks that form the process of developing software.  Brooks becomes a fierce advocate for what he calls “conceptual identity,” a necessary precondition for successful software development that can only come about by the work of a single mind or a small team of likeminded individuals.  Railing against systems designed and developed by large teams or committees, he likens poorly managed software projects resulting “from the separation of design into many tasks done by many men” to “tragically” disjointed European Cathedral designs spawned from construction schedules spanning several generations, where “later builders attempted to ‘improve’ upon the designs of earlier ones.”  

Brooks’ approach – that the best way to manage the software development process is to severely restrict the number of persons involved in its design – became the industry’s standard, and was more or less left unchallenged in the software community until the emergence of the open source model in the early 1990’s.  In a 1997 landmark essay exploring the open source movement, “The Cathedral and the Bazaar,” Eric Raymond challenged the notion that complex software systems “needed to be built like Cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation,” and instead promotes a model “resemble[ing] a great bazaar of differing agendas and approaches.” 

However, it should be noted that Raymond’s concept of the “great bazaar” should not to be confused with the contemporary notion of “crowd sourcing” or other simplistic “free-for-all” distributed labor models.  In his analysis Raymond identifies several specific preconditions necessary for successful implementation of the open source model:

  1. A leader that recognizes good design ideas from others, and possesses the charisma to attract people and keep them interested and engaged with the project’s success.
  2. The presence of a “plausible promise;” that the project will eventually “evolve into something really neat in the foreseeable future.”
  3. Access to a (cheap and easy) medium to exchange and archive knowledge. 
  4. “Openness to the point of promiscuity.” 

Additionally, Raymond identifies several specific “truths” or “lessons” that further inform the success of the open source model:

  1. Contributors for any given project are self-selected.
  2. Every good work of software starts by scratching a developer’s personal itch.
  3. Good programmers know what to write. Great ones know what to rewrite (and reuse).
  4. Given a large enough co-developer base, almost every problem will be characterized quickly and the fix obvious to someone (or put another way, “Given enough eyeballs, all bugs are shallow”).
  5. While one can test, debug and improve [a project] in bazaar style, one should not attempt to originate a project in Bazaar mode.
  6. If you treat your beta-testers as if they’re your most valuable resource, they will respond by becoming your most valuable resource.
  7. Any tool should be useful in the expected way, but a truly great tool lends itself to uses you never expected.

As demonstrated in the following pages, Raymond’s “necessary preconditions” and “truths” to the successful open source model can be similarly leveraged to great affect by the agency attempting to activate the mobile channel on behalf of its clients.


Pros and Cons of the Agency “Cathedral Building”

Before delving into the proposed “Agency Mobile Bazaar” model for the monetization of the channel, one must first examine the current system utilized by media agencies to activate the channel in order to identify why the model is failing to meet industry-wide expectations.   

As defined by the previous section, it is apparent that “Cathedral Building” remains the near universal agency model for the monetization of the mobile channel.  The process basically works as follows:  A “mobile expert” (or in very rare cases, a small group of two or three members) is responsible for keeping up to date with the latest movements of the mobile landscape; distributing a highly condensed version of such information (usually in the form of periodic emails and/or a series of “mobile 101” presentations); and consulting with brand teams for the purposes of creating, selling in, and executing mobile advertising/marketing campaigns.

The model has gained widespread acceptance due to the fact that (at least on the surface) it appears to be an efficient use of agency resources.  Information related to mobility and mobile advertising/marketing is highly specialized, lending credibility to the application of an “information silo” in the form of the “mobile expert.”  Additionally, to date agency revenues related to mobile advertising planning and buying have been nominal at best, significantly restricting the resources available to the agency to put against mobility.  Primarily, it is these key concerns that have given rise to the current system: The emergence of the agency “mobile expert” and the resultant “cathedral building” model.

Unfortunately this model clearly fails to meet the core objective of leveraging an expertise in the mobile channel for the benefit agency clients.  To begin with, the sheer volume of knowledge needed to consistently and competently “stay abreast” of the mobile space, inclusive of the latest research, leading players, usage cases and toolsets is beyond the capabilities of a single individual to follow, much less communicate to a larger group. 

To this add the Sisyphean process of providing “spot” mobile consulting services to a large group of brand teams – all representing a wide range of clients from dissimilar business categories.  The term “spot consulting” is used as (unlike the formal consulting process) the consultant must settle for an intermittent, irregular and at best tangential connection to the core goals, brand objectives and strategies as crafted by the client and agency brand team, with little prospects for a more meaningful bond. 

The Spot Consulting process typically begins with a brief period whereby the brand team briefs the mobile expert on the particular goals of the campaign or plan, followed by a condensed working period where the Mobile Expert crafts a set of mobile recommendations.  It should come as no surprise then that the “spot consultancy” model ultimately yields a consistently low quality of work, as the consultant is operating with severely limited knowledge of the account while operating under appreciable time constraints.  

Beyond this, the process is an overwhelmingly inefficient use of agency resources: Brand teams are forced into woefully incomplete yet wastefully redundant internal briefing cycles with the Mobile Expert, while the Mobile Expert spends an inordinate time attempting to “flash learn” account histories, permanently distracted from his or her of core goal of information-gathering and sharing.  This affect is of course compounded when attempting to scale the model, which ultimately results in even lower quality work, lost revenue opportunities, staff burnout and turnover. 


The Mobile Bazaar Model for Media Agencies

Clearly then a need arises for a scalable agency model for organizational expertise in mobility.  The following “Agency Mobile Bazaar” Model draws upon the lessons learned from the open source software development movement by tapping into the primary resource common to all digital media agencies: A large pool of ambitious, digitally-savvy, young marketing professionals.

At its core, the Agency Mobile Bazaar model uses open source software techniques to “de-silo” mobile expertise and executional competency to arrive at the following desired state:

  1. Like other forms of digital media, mobile media planning, buying and execution occurs at the account team level, by individuals with an intimate understanding of the account; its needs, goals and personality.
  2. The mobile media planning and execution process achieves efficiency and effectiveness through a shared organizational intelligence, populated by an accessible, relevant knowledge base and a talented pool of (internal) mobile media enthusiasts.
  3. High-level mobile strategies, alliances and other corporate-level directives will be driven by an individual or core group of mobile specialists focused on the channel.

To summarize from previous sections, the successful agency model for the monetization of mobility must effectively meet the following core product-specific challenges: marketplace fragmentation, lack of standardization, poor infrastructure and limited channel reach / low perceived value, and labor-specific challenges: maintaining a level of high quality work/output; preserving a market-leading expertise within the mobile channel; limited resource availability; and scalability across multiple geographic regions, brand teams and client business categories.

Recommendations for the implementation of the Mobile Bazaar Model are summarized in the following table: 

Stage Timing Objectives Primary Activities Toolset
Planting 1 – 3 months Educate agency  Educate brand teams PowerPoint
Recruit initial participants Evangelize mobility Excel
  Energize base Email list
  Share learnings (one-to-many)Spot consulting Shared network access
Cultivation 3 – 9 months Establish shared workspace Spot consulting / collaboration Internal Wiki
Build expertise among base Recruit additional participants Internal Blog
Expand participant base Energize base List Serve
Nurture community identity Share learnings (few-to-many) Online Group
Harvest 9 – 18 + months Establish expertise among base Information gathering Internal Wiki
Eliminate most spot consulting Recruit additional participants Internal Blog
Cement community identity Energize base List Serve
  Problem solve (many-to-many) Online Group
  Share learnings (many-to-many)  

Planting Stage.  The first stage of implementation may appear to the casual observer as “Cathedral Building” in that the mobile expert is occupied many of the same tasks, such as meeting with account teams to educate them on mobile advertising, and making ad hoc recommendations to their account teams as requested.  The key difference is that while certainly all of these activities are important, the primary objective of the planting stage is to build interest in / recruit participants for the “Agency Mobile Bazaar.” 

Recruitment can be accomplished in a variety of ways, but in general will leverage the following benefits to the target, who will likely be a junior to mid level member of the digital account team:

  1. Ambition.  Mobile is the future of interactive media; Become an early leader in the medium and your career will rise with the prominence of the channel.
  2. Exhilaration.  Simply put, Mobile is cool.  iPhones are cool.  GPS targeting is cool.  QR codes are cool.  Pattern recognition is cool.  Mixed reality gaming is cool.  IVR, SMS and even Bluetooth can be cool.  This is the reason you’re in digital media in the first place, and not outdoor or print, right?

The recruitment process will begin in person (with the individual digital client teams) at the “Mobile 101 / State of Mobile” presentation, where the mobile advertising space will be positioned in the most hopeful and exciting of lenses.  This is not to say that client teams will be fed “marketplace hype” -quite to the contrary.  Client teams will be presented with a realistic outlook of the channel, the potential which will nearly speak for itself (a knowledgeable and charismatic presenter is also a requirement). 

Near the end of the “Mobile 101 / State of Mobile” presentation the Mobile Expert will solicit members for this new group of “Agency Mobilists” (working title), the sole “membership” requirement of which will be a genuine interest in mobile marketing –and nothing more.  Group Directors may be consulted in advance of the meeting for recommendations on possible “best fits” within their team, as well as for any additional help in recruitment process.  The primary goal for the Planting Stage is to recruit one Agency Mobilist per digital account team.

During this phase the Mobile Bazaar community will begin to take shape by a variety of tactics, all of which will rely on decidedly low tech and/or Web 1.0 tactics: Bi-weekly emails of “cool” mobile marketing news, interesting case studies (nothing boring!); and access to key resources such as mobile network reach, targeting and contact information and presentations (shared via access to a shared network partition).  Community feedback for this early phase will be passively solicited in the form of manual techniques, such as email.  From its earliest possible point, any “official” community correspondence and/or language will be inclusive, supportive and decidedly casual.

Additionally, “spot mobile consultancy sessions” will be still be utilized during this phase to support the client teams, but executed quite differently than previous models, and with two added goals: The passing of mobile expertise, and the forming of community bonds.  Whereas in the past the mobile expert would, in perfect “Cathedral Building” style, work up a set of recommendations more or less in isolation from the client team, in the Bazaar model the mobile expert would work directly with the client team’s resident Mobilist, providing insight, support, resources and recommendations, but will refrain from actually building the proposal proper.  

By encouraging, reviewing, supervising and supporting (but not actually authoring) the mobile recommendation set, the mobile expert will in affect transfer ownership of the work to the Mobilist, instilling all the pride and satisfaction that comes with possession.  If no Mobilist is yet present in the client team in question the consultancy session itself becomes another recruitment opportunity.  Ultimately this process – that of planning, selling-in and executing of mobile strategies and campaigns – will be shared with the group via the aforementioned communication channels.  Community feedback will be encouraged but never expected, observing Raymond’s rule of self-selected participation.


Cultivation Stage.  The intermediate phase of the implementation plan is most prominently marked by a marked upgrade in collaboration tools.  Shared network drives and other manual communications channels such as PowerPoint and Excel will be de-emphasized in favor of more effective tools. While the email list will continue to be important, more often than not it will be utilized to point traffic to a relevant internal blog posting or internal wiki article.   Information from the legacy formats (such as an Excel database housing data on mobile ad networks, for example) will be transferred to the new, more accessible (and malleable) mediums.

The internal blog, published primarily by the Mobile Expert, will assume the ideological center of the Mobilist community, providing an essential axis for the group and the expression of its personality.  Postings will reference / drive traffic to relevant internal wiki articles; highlight any prominent forum chains, internal case studies or ongoing projects; praise good works and humbly revel in the satisfaction and exclusivity of expertise. 

That said, the blog is by definition a “one (or few)-to-many” medium, and (even with the ability for users to post comments on existing posts) as such is not a true “community channel,” hence the need for the more robust communication channels such as the wiki, list serve and forum.  While early participation in these channels will likely be limited to the Mobile Expert and a small group of like-minded emerging media specialists inside the agency, it is eventually through these channels that greater Mobilist community will begin to share information about working projects, learnings, vendors and other points of interest – soliciting others for help on a particular challenge or simply bragging about the results from a recent campaign.

The legacy Spot Consulting process will gradually give more ground to the “Spot Collaboration” method, where the Mobile Expert’s advice is supplemented by that of the community.   As with the previous stage, the Mobilist – not the Community and certainly not the Mobile Expert – will assume an ownership position of any and all mobile campaigns, as the client teams become the driver of mobile activity within the agency.   It is expected that the Mobile Expert will be consulting on key clients, but the actual need for his or her services will vary considerably from account to account.

The Mobile Expert can now begin to focus more effort on identifying corporate partnerships and alliances advantageous to the agency, and researching new mobile planning and execution tools as they emerge in the marketplace. To this the leader’s role is to keep the base engaged by continuing to introduce exciting and entertaining innovations from both the corporate level and from the field, promoting any local “mixers” or industry events of interest, and maintaining an inclusive environment where ideas and new thinking are valued.  Recruitment efforts will continue but will not be given the same emphasis as in the Planting Stage, as it is expected that organic growth will offset any natural turnover in the system.


Harvest Stage.  It is the final stage of the implementation plan where the agency will be able to fully reap the benefits of the Bazaar.  At this point the Mobile Expert’s relative contribution to the community will appreciably diminish, as the cooperative spirit of the group increasingly asserts its dominance.   The leader will continue to participate in community-building efforts via blogging and other means, but will by no means overshadow the conversation.  Mobilist Bazaar itself, via shared knowledge and experience, the will eventually become “the mobile expert” in and of itself, collectively superseding the limited resources that the Mobile Expert can bring to bear on any single project or campaign.

As Mobilists begin to plan and execute mobile advertising campaigns independent of the Mobile Expert, the latter can focus more fully on high-level corporate initiatives.  Spot Collaboration activities by the Mobile Expert will no doubt continue on key accounts, but in a greatly reduced volume and at a much higher level – significantly increasing the quality and effectiveness of such recommendations.With this process, Mobile expertise is continually distributed to the client teams, as the model reaches its desired state. 




If effectively applied, the open source model can provide a solution to the vexing “chicken and egg” scenario agencies face when attempting monetization of the mobile channel. 

In the software industry the model has already produced brilliant and wildly successful software applications such as Apache Web Server, mySQL, Firefox and WordPress – applications built by a community of volunteers that have come to dominate, outperform and/or out-innovate competing products produced by many teams of full time programmers from the likes of Oracle, Sun, and a host of other industry giants.

Likewise, industries such as Medicine, Law and Finance which face a similar set of challenges (high-complexity of tasks, communication failure, poor distribution of labor, and severe time and cost constraints) are also translating the open source model as a scalable solution to meet their seemingly unique needs.As with most successes, the key to ours will ultimately lie in its execution.  Beyond the few rudimentary software needs detailed above (the best of which are, ironically, all open-source / free products), the most crucial component that ultimately drives the success of any open source project is the Leader’s ability to fully embody and express the ideals of successful open source community: An egoless enthusiasm for the space, a full-throated transparency of thought and of action, and willingness to value and reward the thinking of others.  

It is through the positive expression of these ideals that the community takes shape, grows, and in the end, achieves far more than any individual or formally organized group in terms of the quality and quantity of thought, insight, knowledge and productive work.