Before I start, I freely admit to it being ludicrous that we will ever be predict where a team will end up in terms of the practices it deems useful and the impact of the constraints of its context on how it operates. The same is likely true of organizations tackling a desire to scale their Agile practice. Despite Agile passing the 10 year mark it seems like we will never have a prescribed trajectory for the evolution of Agile teams and if we did wouldn’t that take the fun out of being on a team? Learning and growing in a way in which the mutual needs of the team members were attended to, as well as the needs of the folks the team was serving.
Agility of the Ages
I however can’t escape my fascination with the question of what part of our psychi Agile is feeding off when we look to it to satisfy needs as to how work is done in a team environment. If it encourages an environment where we can find our state of Flow, is it an echo of a time when we worked in small teams for so long that it became a part of us? Maybe Agile is not a new mindset but a return to how we had been building complex products successfully, but had forgotten.
Popular articles on the Spotify team model, referencing charters and guilds, made me wonder if the descriptions of guilds bear any resemblance to Agile teams. I started reading “Guilds, Innovation and the European Economy 1400-1800”, a series of essays and papers edited by S.R. Epstein and Maarten Prak. I don’t claim to have found any any direct correlation but there does seem a number of parallels around how guilds operated within an array of socio-economic contexts. These contexts varied by whether the power lay with merchants, guild masters or the politics of the town. They shaped the myriad of ways that guilds operated within the trade based economy of the Middle Ages. When I get through more of the book I’d like to pull out those parallels. For now I suggest that in medieval times, guilds were often viewed as the corporate components to achieving scale, much as teams are viewed in SAFe
The benefits of guilds are not dissimilar to those of teams operating within SAFe – to reduce agency costs in collections of guilds often considered an early form of firm or corporation. Typical agency costs include those incurred when the buyer or seller did not fully understand the trading partner’s situation. Asymmetry of information was caused by merchants not knowing the quality of goods or producers not knowing the financial viability of a merchant. The merchant’s concerns were often solved by installing oversees or product managers. This is referred to as Delegated Monitoring. I see this as not unlike the role of Product Owner being embedded in the team as a proxy for the true customer’s needs.
As products became more sophisticated, being comprised for the work of different skills or components, guilds needed to become vertically integrated. Additional agency costs were generated each time a buyer needed to stop and inspect the goods. The larger the enterprise, the more points to unpack, inspect and repack goods. The inspection costs could be avoided if the intermediate goods became standardized and if the practices of the guild are such that trust through reputation can be established. The latter resulted in the creation of trademarks applied to finished goods. I find the idea of our build artifacts analogous to standardized products. The application of automated code inspection and testing establishes a similar notion to quality inspections.
Two Path to Scale
The reason the book I am reading was written was to pose a counter to the complaints often made of guilds. That they stifled innovation and in doing so became the reason for their own extinction as the industrial age and capitalism superseded the age of guilds.
One chapter I recently finished addressed how guilds arranged themselves within larger economies of scale and it made me think of the current debates around the merits of SAFe. I have been working on a program which purports to be leveraging SAFe for two years. I think it unfair to lay the challenges of the program at the feet of the framework whether you are a proponent or not. However I do wonder whether the time spent by us trying to get SAFe “right” would not have been better spent organically growing something that “made sense” with the constraints we operate within.
In the specific examination of scaled guild-based economies the pattern emerges of essentially two paths to scale. One driven by a merchant focused view and the other from the producer focused view. The idea of the two paths was an elaboration formed by Maurice Dobb and Kohachiro Takahashi on work originally done by Karl Marx. He promoted the idea of true revolution in industry coming when entrepreneurs were developed from the ranks of the tradesmen. This view of the transition from feudalism, Takahashi argues, was more of a western phenomenon. The idea of the merchant becoming the producer he declares evolved in Eastern and Asian economies. Both were present in the age of western guilds. He explains that the two paths are really not different attempts to solve the same problem. They emerged as a consequence of the tension between of the social groups: guilds, merchants and political bodies. The needs of affluent masters seldom aligned with eager merchants. In other words, there was no best way to scale.
Merchant as Entrepreneur
If merchants were driving the aggregation of guilds into an a proto value stream they tended not to worry about the evolution of a systemically optimal engine of innovation and practice. They looked to leverage existing economic capabilities. They would often buy raw materials and sell them to the guilds at a discount in a system called “putting out”. It bootstrapped the relationship. The guilds would then sell the finished good back to the merchant thereby recovering the costs of materials and making a profit. The merchant in turn would sell to their markets. Merchants tended to prefer this system which created a dependence on gain from arbitrage (“buy low to sell high”) cycling capital through the parts of the economy they understood and could control. This arrangement mimics how our SAFe program has worked with the portfolio layer, where raw ideas in the forms of investments, are handed off for elaboration managed by program layer.
The program layer in medieval times could be organized by guild masters or by merchants operating as entrepreneurs. Using their respective understanding of the extended production process, they would arrange expansive contract agreements between component guilds to string together a distributed predecessor to the industrial production line. This middle layer would use an early form of subcontracting with smaller guilds or firms to get the capacity needed. This layer seems most like the Scrum/XP teams of SAFe. The emphasis then is more on flow but less on what we might call becoming more Agile. There is less focus from the merchant on developing an adaptable economic instrument that could handle changing markets while remaining intact. A merchant would be just as likely to form a new enterprise as transition an existing one.
Producer as Merchant
This path allows for a systemic view of the role of the aspiring master-entrepreneur. Straddling the layers, the master need only leverage subcontracting arrangements for capacity. The book makes the point that in this age, innovation was not impeded wholesale by guilds. Rather emphasis was given to product innovation over process innovation. This was true regardless of which of the two paths were in play. The degree depended on the influence of the masters involved which in this path was more significant. Innovation tended to come about with tacit knowledge gained by learning from doing and using. Not unlike the notion of retrospectives and gemba.
With more complete oversight of production, a master could make the management of agency costs, that aligned with our notion if quality, compatible with product innovation.
The books supports the idea that scaled guild-based economies could be innovative if the masters were empowered to be engaged in the subcontracting arrangements as well as having full exposure to the market needs the corporation served.
Are There Two Paths to Scaling Agile?
To bring this forward to the current discussions of the validity of SAFe I would ask is there a similar analogy available as an alternative? If SAFe aligns more to the merchant driven model then is there an alternative in which we bring the masters of the software crafts into the light of the ultimate customer. In doing so do we reconnect with what has in the past been central to our ability to scale? Can we then hold onto intrinsic quality while innovating what I call extrinsic quality – the degree to which the product meets the needs of the ultimate customer?
Maybe only time will tell. Maybe we need four hundred years of time to tell.
Guilds, Innovation and the European Economy, 1400-1800 by S. R. Epstein et al. http://amzn.com/0521153913
The Transition from Feudalism to Capitalism: A Contribution to the Sweezy-Dobb Controversy : H. K. Takahashi and Henry F. Mins Science & Society Vol. 16, No. 4 (Fall, 1952), pp. 313-345 :http://www.jstor.org/stable/40400147