Feedback and the Story of Phineas Guage
We started with the highest of hopes. We were destined for great things. We were going to plow that green field. We weren’t going to be bound to any rigid process. I’ll spare you the details, because if you stop and reflect for just a moment on your own experience with these sorts of things, you probably already known where we were headed.
It’s true that our ad-hoc Agile had devolved into a patchwork of poorly defined initiatives, unrealistic expectations and missed deadlines. Considering our weekly burn rate, we knew it couldn’t go on like that forever, but somehow it seemed there would always be time to reform.
For the longest time, management made no obvious attempt to staunch the bleeding, but all the while, they were preparing their move. With the kind of results we were giving them, they had little choice but to act. The problem was seen as lack of structure, and so structure, plenty of it, was seen as the solution.
They brought in Change Agent contractors to train Executives, Managers and Leaders on how to identify Value Streams and to assemble Agile Release Trains. Implementation plans were drawn up. Release Train schedules were prepared. Coaches and servant leaders indoctrinated the willing and the weak alike. Outposts of the new culture were established at strategic points throughout the organization.
One could recognize the zealots by their excited muffled whispers of Weighted Shortest Job First and Program Increment Milestones. We listened in awe as management called for a Spike into the Spanning Pallet, waxing eloquent on the critical importance of the Architectural Runway. When it got to the part about leveraging the Solutions Backlog to get everyone on the Agile Train, I had to pinch myself to make sure it wasn’t just a bad dream.
“In days of old, enterprise software teams wandered aimlessly for many moons and even the IT chieftain knew not of a safe haven. Then Schwaber and Sutherland came to lead the tribes out of the desert, to a new land of Milk and Scrum."
“In this land of dogged persistence and incremental accomplishment, the ancient muddy ritual of Rugby was consecrated to a new era of software development dedicated to moving the ball down the field and never giving up ground."”
The mere suggestion of Kanban was understood as disloyal subversion. The pressure was on to be seen as a team player. That line from Hamilton came to mind: “Talk less, smile more, don’t let them know what you’re against or what you’re for”. The hegemony of SAFe culture was uncontested.
I retreated into my requirements hide-y-hole, as Masters fresh out of Scrum school preached the gospel of “sprint commitments” and the reeducation programs that awaited those who couldn’t live up to their Sprintly pledges. The hell fire of burn-downs became the daily catechism, and velocity was measured with all of the precision and reason of the alchemists of old in pursuit of the secret to forging gold from mud and blood.
Certified SAFe Change Agents came as Apostles of a great and undisputed faith; speaking of the benefits that would accrue to those who acted according to canonical practice, — of a day when Agile trains would run on time, and a future at astonishing scale.
The way ahead would be a road paved in best practices.
Meanwhile the teams continued to crank out code. Pull Requests are the fabled land where theory and practice meet for lunch.
_Hey! Long time no see! How have you been? How many little Pull Requests do you have underfoot these days? Really! That many! _
When I bring up the topic of some gnarly dependency that’s blocking a release; I might as well have invited him to diaper changing party.
_ … oh my, how time flies! … I’ve got a two-o’clock, but really we should do this more often!_
And so theory goes off to catch a train.
Say what you will about the fine art of software process engineering, but you can’t get away from fact that it all comes down to releases. Releases are made up of Pull Requests, and Pull Requests are sets of commits. In the end, it all comes down to the code. It brings to mind those immortal words
“In theory, there is no difference between theory and practice, but in practice, there is.”
In practice, it’s the code that’s released to production that matters. It’s called a commit for a reason.
There was a wide gulf between the land where the Release Trains were being planned, and the Pull Request Edge, where code was being loaded on boxcars. More often than not, they seemed to be trains headed in different directions.
A path forward came in an unexpected way. A senior manager took me aside for ideas on what to do about the wreckage that our project had become. There was a plenty of low hanging fruit, but thankfully I had the presence of mind to answer simply: “nothing”, recalling “start from where you are”. My manager thought I was being sarcastic, but for once, I was in earnest.
That’s when the fun began. At first we had the nearly irresistible impulse to design the process to what we imagined it should be, but we recognized that trying to engineer an outcome would likely meet stiff opposition. So we soldiered on, visualizing the work and focusing our inquiry on what was impeding the flow. Once, when we began to suspect that the root of the problem as a lack of systems thinking, we looked at our problem through the lens of the Kanban Maturity Model.
The KMM makes no presumption that intervention is somehow always a good thing. You simply practice Kanban, visualizing the work, to help insure that you actually understand what’s going on; always a good idea, then you limit the work in progress in order to drive the constraints to the surface. Only then do you consider policies and practice changes, and only then as a response to the problems you actually have. The model helps in making appropriate choices.
You could think of the KMM as a “systematic protocol to determine when to intervene and when to leave systems alone.” Imagine that, the idea that sometimes doing nothing might be your best option. The KMM doesn’t insist on piling on practices but rather provides meaningful guidance on how a company might mature incrementally.
At that point, my colleague asked what Maturity Level he thought we were at.
Instead of answering the question, I asked a few of the standard questions:
Do customers perceive service delivery as unreliable? Is there an observable lack of alignment among teams? Is work pushed into the system? Is the process stream is overloaded?
The answers to these and other questions indicated that we were still at the beginning our our journey, certainly some things in our favor, but subject of the occasional regression under the stress of deliveries. That led to the question, what practices should we be looking to adopt?
Like other Maturity Models, the KMM helps guide in practice adoption but with the critical distinction is that practice adoption is not a measure of success.
The only real measure of success of an organization utilizing the KMM should be productivity. Many technology managers use the word “productivity” rather loosely; often they really mean some local optimization along the lines of keeping everyone busy, or perhaps they want to emphasize the importance of deadlines by equating productivity with performing according to their GANTT chart. Rarely do we find the word productivity used in a context that is accompanied by Andy Grove’s formulation of “…as measured by…” or anything, for that matter, remotely in the direction of specificity.
Kanban measures productivity in terms of throughput, and, once you have a stable system, throughput can be measured reliably using Little’s Law.
We can argue about the meaning of productivity in terms of additional measurements of the business value of delivered work, but as Eliyahu Goldratt pointed out in his critique of the Balanced Scorecard, there’s virtue in simplicity. Throughput doesn’t answer all our questions about business value, but it is a sufficient metric for the context of evaluating the relationship of practices with productivity.
My manager liked the part about Little’s Law, the whole thing made me come off as some kind of expert, very cool math that you can count on hardly anybody knowing much about.
John von Neumann said that, if you want to win a debate, you should mention entropy, because people won’t want to admit that they’re not sure what it means.
Imagine you’re loosing an argument with someone from the Project Management Office. Not hard to do. You just let them get their point off and act as it you’re agreeing with them, and then, at the last minute, like it’s an afterthought, ask if they remembered to factor entropy into the estimate. That’ll slow ‘em down.
At this point in the conversation I should just fess up that I’m very much at risk of becoming a victim of my own expertise.
“Dearly beloved, we are gathered here together today to lay to rest any chance we had to actually understand the system, and content ourselves to bits and pieces of some crafty knowledge."
Experts are best at planning their own funerals.
We spend our whole lives trying to become experts, only to fall on our own sword. We read books and attend the best conferences, like this one. While normal people talk about sports or how to barbecue a goat, we debate the finer points of story mapping strategies. We study hard, even reading the footnotes. We become experts. All this is necessary and good, so long as we don’t fall into the trap of thinking that having expert knowledge is the same thing as knowing what to do with it.
Everyone else seems to know what to do with your expertise: you’re supposed to solve their problems; to recommend the best solution or the most likely outcome. You’re expected to act as a sort of problem-solution match maker. Many consultants just assume that that’s what the game is all about; the Harry Potter Charm school model of consultancy, where missed deadlines can be remedied with the “Acceleratium” charm, or quality issues vanquished with a “BestPractium” spell.
We know that you’re not one of those coaches selling solutions out of a Sear’s and Roebuck style catalog of “Best Practices”; that you’re the level headed type, making a serious effort at using your intelligence to apply hard-earned knowledge to the problem at hand. Considering everything that you put into becoming an expert, it’s only natural that to have the nearly irresistible instinct to respond to the challenge of a problem with a definite solution. What kind of expert are you if you can’t do that?
Let’s take a look at how expert solutioning thing works. Presented with a problem, you make a scan of your extensive experience, looking for patterns that seem match the nature of the problem. We do this intuitional pattern matching in an instant, setting the context to all of the critical thinking that follows. Most of what we think of as expert analysis is really just giving a story to our intuition. For a deep dive on what’s going on here, see Danial Kanaman’s “Thinking Fast and Slow”
Suffice it to say that chimpanzees throwing darts do surprisingly well against experts at the top of their field. Becoming more expert doesn’t seem to help that much. If chimps could learn to show up to meetings on time, they could turn out to be formidable competitors to many Agile coaches, especially when you consider their discounted billing rate. Although this may seem demoralizing at first, there is no need to despair. The problem isn’t that expertise is valueless, it’s just that we apply it in a somewhat ineffective way.
Let’s look at the question from a different perspective.
In 1932 the American Geneticist Sewall Wright developed a concept of studying evolution by visualizing the distribution of fitness values. This mental model describes an immense grid called a fitness landscape, with each cell containing some random combination of the species attributes. Most of them will be totally unworkable, some will be viable although destined for an early exit from the gene pool. Finally, there will be clusters of grids occupied by species well suited to the environment.
This two dimensional grid concept is an improvement over the usual expert strategy of picking winners, which amounts to picking one of the grids without much in the way of awareness of the whole landscape. but Wright took it a step further, transforming the fitness landscape into a 3 dimensional grid, with the really awful solutions rendered as troughs and the viable solutions being peaks.
Now add the dimension of time. Adding the temporal axis results undulating peaks and valleys; as the conditions of the ideal environment, so do the attributes that correspond to fitness. The winning grid square selected through expert analysis today becomes tomorrows tough. It turns out the the dartboard is more like a magically morphing target.
Now think of a problem you’re asked to solve. Consider all of the combinations of all of the variables and possible system states. That’s the fitness landscape where you have to survive and thrive. Our ability to show our clients around the peaks and valleys of their own fitness landscape is a good opportunity to put some distance between ourselves and the dart throwing chimps. Any expert can land a dart on the board, but it takes someone with a good sense of direction to navigate the fitness landscape, someone who can think in terms of the system as a whole.
This mental model broadens our view to understand that systems thinking is what sets us apart. Once we grasp the complexity of our problem space, we can see the deficiency of trying to pick winners through expert analysis. We need a different approach.
The fitness landscape gives some sense of the journey ahead, but we still face the problem of inertia.
“Start from where you are” is not just a slogan. What’s the alternative? Start from where you aren’t? I suppose that’s the whole idea of Sprint Zero, to emphasize that wherever it is that you are, you’re going to have to get down to the bus stop on time to catch the Sprint 1 Line. Don’t be late!
You’ll never miss the bus when you start from where you are, because you’re already there. Just keep in mind that before it was a starting point, it was the end point of whatever came before, and every fresh start is going to include some small measure of resistance to change.
Strategies for overcoming inertia are just as important as expert knowledge of the problem domain.
The parable of Buridan’s Ass tells of the donkey standing exactly half way between water and food unable to decide which way to go. Although his inability to choose will be the death of him, you could stay that the donkey is in a stable state. We tend to think of stability as our goal, but sometimes it is what stands in the way of improvement. While our donkey domain expert is off building a Monte Carlo simulation to determine the probability of the donkey dying of thirst of starvation, someone comes up from behind and whacks the donkey with a 2X4, a black swan event from the point of view of the donkey expert. We don’t know which way the donkey is going to go. It turns out that we don’t care, often just setting things in motion is more important than knowing what the outcome will be.
The metaphor of the Grey Squirrel sheds a little more light on our question. Although the Duke of Bedford first brought Grey Squirrels to England in 1870, the reign of the greys did not really begin until 1911, when the Duke of Burmingham delighted guests at the wedding of some by now long forgotten aristocrats with a pair American Grey Squirrels, Sciurus Carolinensis. Seeing the liberated greys scamper off into the woodlands of the Conventry Park was a charming symbol of the happiness in store for the newlyweds but distasteful and distressing for the natives, Sciurus Vulgaris, the Red Squirrels watching from tree perches above. Something was set in motion on that day that eventually caused the Reds to abandon the English forests. Today you’d have to travel as far as Wales or Scotland to find a family of Reds.
The story of how the Grey Squirrels came to prevail over the natives is often told in the Kanban community as a metaphor, of incremental evolutionary adaptation; that even a seemingly insignificant change can bring far reaching results when it proves to be better adapted to the problems posed by the environment. But a closer look at social and physiological dynamics of squirrels can lead us to be better understanding to what this metaphor is about: the role of stressors in the fitness landscape.
It’s been more than a century since that fateful wedding party, and the internet is full of inflammatory and divisive theories about why red squirrels are exiles today from their former homeland. There are those who spew their digital invective over Grey Squirrel aggressiveness, while remaining silent about reports of equally belligerent Reds.
Some people cite scatter hoarding, storing their nuts in multiple places, as the key advantage of the the Greys over the Reds, who go in for the local optimization of a single storehouse. It does seem to have the same class of advantage as index funds, but as compelling as this theory is, it is doubtful, because the Reds don’t appear to be starving.
If you meet some old timer from the Scottish Highlands, as soon as the topic of Squirrels comes up, (which won’t take long) before you know it, he’ll be off on a jag about a pox on the Red Squirrel and the disease bearing Americans and how he’d shoot one as soon as look at ‘em. If you do run across one of these old timers, you might as well settle in for a pint, because there’s something about Scotts and Squirrels, where once they get started on the topic, there’s no stoping them.
Some people seem to imply the the simple fact that the Carolinensis are of American origin to be a sufficient explanation for their dominance. It don’t know whether to take that to mean that they’re better or just bullies. As you might imagine, most of these theories are simply lacking in rigor.
It is interesting to note that Reds and Greys are known to cohabitate the same woodlands without much in the way of observable conflict. Researchers Luc Wauters and John Grunell spent two years radio tracking the critters, publishing their results in a 1999 issue of “Ethnology Journal”. They concluded that “there as no more aggressive encounters between squirrels at the Red-Grey site than at the Red-only site.” In most cases, the two species seem to ignore one another. Internet trolls conveniently overlook such details as they spin their Grey Squirrel conspiracy theories.
In “The True History of Grey Squirrels in Britain” John Bryant notes that “It has been alleged that the (pox) disease was transmitted by Greys, who are immune to the virus. In fact, out of 44 districts where red squirrels were affected over 20 years, only four had grey squirrels present!” So the Grey Squirrels as disease carriers turns out to be a red herring. So what is this Grey Squirrel thing all about, anyway?
Recently, a distinguished team of researchers provided the kind of the hard data that fair minded people have waiting for in this question, in the form of a peer-reviewed paper published in the scientific “Journal of Animal Ecology” titled: “Stress in biological invasions: Introduced invasive grey squirrels increase physiological stress in native Eurasian red squirrels.”
These intrepid scholars put some rigor to the question by gathering Red Squirrel scat. In their own words, they (quote) “…extracted glucocorticoid metabolites from fecal samples to measure whether the presence of the invasive species causes an increase in physiological stress in individuals of the native species.”
Glucocortiocoid is a part of a feedback mechanism which serves to reduce or suppress the mamallian immune system. You’ve got some too. In a healthy subject, it’s just another hormone that shows up at the right time to keep the system in balance. In pharmacology, it can do the trick to put an overactive immune system response back in its place. For some reason, it is overproduced in stressful situations, so if you’ve ever wondered why stress can be harmful to your health, Glucocortiocoid could be a good place to look. Too much of the stuff floating around, and you’re more vulnerable to all of the things that your immune system is there to protect you from.
So what conclusions did the scientists reach about these formerly contented English Red Squirrels? The presence of the stress hormone Glucocortiocoid was found to be three times higher in Red Squirrels cohabitating woodlands with greys as those who still lived in forests as yet unsullied by the American Carolinensis. The grey squirrels aren’t stealing nuts from the natives, they’re just making them nervous, and effective stressor.
The stress that red squirrels feel in the presense of greys results in increased levels of Glucocortiocoid, which supresses the immune system, making them more susceptible to disease, such as squirrel pox. What British Red Squirrels need is not the extermination of their American cousins, but simply effective coaching to adopt practices appropriate to their maturity level which will alleviate their stress.
Red Squirrels are clearly in ML-Zero “Oblivious mode”, and would benefit greatly by learning a few effective strategies, such as learning to engage in personal reflection, and defining work types based on the nature of tasks. These practices as proven to reduced stress, which will bring down Glucocortiocoid levels and strengthen the immune system. With proper coaching, Red Squirrels will have , so please, put away your shotguns and use your coaching skills to get the Reds back in the game.
Had the Duke of Burmingham consulted with a wildlife professional on how to evict Red Squirrels from his estate, you can bet that the expert solution to the Duke’s problem would have involved quite a bit more effort than lifting the latch on the cage of a couple of mail-order rodents. Of course the Duke didn’t think of his Red Squirrels as a problem needing solving at the time, but having the presence of mind to spin a retrospective narrative to lay claim some irrefutable outcome, thats the stuff that consulting legends are made of. Any chimp might beat you at the dartboard, but you can always come out ahead by telling the story of how the dart landed just where you intended.
Armed with the secret weapon of retrospective narrative, we can apply the solution of the donkey’s dilemma to nearly any problem. A wack with a 2 by 4, a pair of grey squirrels, whatever it takes to put inertia under stress, because the problem isn’t really the problem, it’s the stasis that resists change that is usually the tougher nut to crack. A system which isn’t under stress is one that is stable. Stability doesn’t much like disruption, even when that stable state is sub-optimum. As Forrest Gump said, sub-optimim is as sub-optimim does. The key is to not let sub-optimum settle in and get comfortable, but rather to keep things in motion. No need to worry about where things will land, you can always come up with story later to explain how you were headed where you ended up all along.
If you feel that you don’t have time for your squirrels to do their magic, you can always put on your expert hat and gown and just tell people what to do. Most organizations will expect nothing less. The fact that they expect it doesn’t mean that they’ll go along with it. Colleagues from all ranks of the organization will sing Kumbaya around the campfire with you all night long, but then cross shields in defense of the status quo in the morning. They’ll be so busy guarding the ramparts that they won’t even notice grey squirrel clambering up the wall.
It’s hard to get anything done in the face of inertia, so you can see why the metaphor of the Grey Squirrel is so powerful. But don’t get the idea introducing stressors is the whole game. When your problem is inertia, introduce a stressor to get things in motion, stir things up, trigger a response. When there’s a positive response seeking stability; that’s when it’s time to consolidate your gains.
The KMM gives us a mental model for these two modes: transitional vs consolidation phases. Transitional to get things in motion and then consolidation around improvements. This idea of phases is not the same as the whatever cadence you’ve arrived at, it is more a reflection of dynamics of inertia and stability. So long as inertia is blocking improvement, you should consider that you’re in a transitional phase. People that are really good to resisting change will also be good at dragging out transitional phases. Some things just take time. The Grey Squirrels didn’t conquer the rainy island in a fortnight.
The KMM gives us another a mental model to help sort out which practices work best in transition phases and which are suited to consolation phases. These are normative vs structural changes.
A normative change is a revision of policy or procedure which doesn’t disrupt social structure, thus is less likely to meet resistance and more likely to build trust in the idea that change can lead to improvement. In contrast, structural changes imply the need for some adaptation to roles, responsibilities or status of team members, hence less likely to succeed without the buy-in from those who are affected. Buy-in is earned by building trust. Building trust provides the foundation to allow team maturity to grow through self reinforcing dynamics of the phases. Team maturity is measured as increased productivity.
The interesting thing we found is that many of the same dynamics apply to any team, including Scrum teams. The concepts of pairing transitional vs consolidation phases with normative vs structural practices is a powerful model for moving things in the right direction. Of course the specific practices of the KMM will be out of reach for Scrum teams, who trade off a focus on managing constraints for the lump-sum constraint of the timebox. It’s a lower fidelity operation, but as Donald Rumsfeld said, you got to war with the army you have.
Whatever the specific practices you’re considering, it still makes sense to ask if you should be focusing on moving the inertia needle, or if the time has come to try to consolidate gains. That’ll provide the context for a discussion of the normative vs structural implications of any particular practice proposal.
For example, when determining when to attempt a move from a transitional phase to the consolidation phase in Kanban, you can design pull policies and other board attributes to visualize the signals. The turning point will be right there on the card wall, getting rid of a lot of guess work. When working with Scrum boards without explicit WIP limits and queue management, you’ll have to work harder to observe other telltale signs.
Thanks to the intrepid researchers from the Animal Ecology community, we now have some hard science to work with to understand how to read the dynamics of stressors.
The secret of knowing when to pivot from a transitional to consolidation phase depends on getting feedback that indicates that your stressors are working. This has been the secret sauce of the top coaches, but now that we know about glucocorticoid, every coach can pivot like a pro. Saving thousands of hours of difficult and expensive trial and error, we now use ScrumScat, the mobile phone app which provides realtime readings of glucocorticoid metabolite from Scrum Master scat. Let the grey squirrel loose and set the trigger levels in the app. When Scrum Master glucocorticoid spikes, you’ll know that it’s time to pivot to consolidation phase, taking the guesswork out of one of the most treacherous decisions faced by Agile coaches.
You can see by now that the problem of getting the right feedback is at the center of everything, so it should come as no surprise when we encounter the feedback problem in scaling frameworks. We’re going to have to get a lot of different feedback mechanisms right before the Agile Trains will all be running on time.
When you bring up the topic of feedback in the Agile community, people seem to think along the lines of the kind of thing that happens in a retrospective session. The kind of information gathered in this way is valuable and important, and it’s certainly an improvement over the kind of insensitivity that was the norm in pre Agile IT departments, but there are other kinds of feedback as well. The Profit and Loss statement, for example. The P&L seems to be considered a bit of a relic in the age of Agile, but it has stood the test of time as a valuable feedback mechanism, especially when avoiding bankruptcy is one of your KPIs.
Something about Enterprise Digital Transformation initiatives has always struck me as odd, with everyone looking much like they did before, but acting in ways inconsistent with their former decorum.
I gained some insight into this dynamic through the story of Phinias Gage, retold by Tim Harford in his book “Adapt”. About 20 years before the American Civil War, Phinias was blasting rock for the Rutland and Burlington Railroad, just south of Cavendish, Vermont. Phinias was tamping down blasting power in a rock bore when the metal tamping rod generated a spark which ignited the explosive dust. At the moment when the powder ignited, Phineas was just turning to talk to his crew behind him, bringing his head inadvertently in line of the trajectory upon which the rod was about to embark. The steel tamping rod was a little over an inch in diameter and just over 3 feet long. The steel rod entered Phinias’s open mouth, passed behind his left eye, clear through the left frontal lobe his brain, exiting out of the top of his skull, landing some 80 feet away.
The unfortunate Mr. Guage had the remarkable fortune to not only survive, but remained conscious through the ordeal and was able to explain the circumstances of the accident to the attending physician a half our later, while the good doctor stared at disbelief at the gaping hole in the man’s skull. Although the steel rod carried off a not entirely insignificant portion of Phinias’ brain, he recovered and went back to work, although not for the railroad. He traveled to Chile and took a job as a stage coach driver on the Valpariso-Santiago line, evidence that he continued in life as more than an invalid. Other than being blind in the left eye, he looked like the same man as before. But there was some distinctive change in behavior. Before he had been known as a reasonable man, but Phinias missing a decent size chunk of his brain was often belligerent and indifferent to social norms that he had previously respected. In short, while he retained nearly all of the abilities of his former self, he seemed to have lost much of his ability to process feedback.
When enterprises embark upon a Digital Transformation odyssey, it sometimes seems as if they suffer a steel rod blasted through the corporate cranium, taking out the part of the brain that processes feedback about profit and loss, while leaving everything else in tact. Whereas before we knew vigilant guardians of the balance sheet, we find a purveyors of retrospective narrative, turtles all the way down.
If an Agile Transformation Initiative is going to improve our ability to deliver value to the business, then we should be able to tell the story of that value in the P&L. It’s true that, under the best of circumstances, the P&L traditionally aggregates cost accounting metrics, while being blind to the value delivered by productivity initiatives. These shortcomings of tradition cost accounting have been taken by Agile advocates as license to just tell a good story in place of doing the hard work of verifiable quantitative analysis.
I understand that it’s not an easy problem, but papering over the misalignment between Agile initiatives and the P&L is a lot like Phineas without the feedback part of the brain. You can count on him showing up to work, but you might wonder what he’s capable of next.
It’s only reasonable the the budget is the voice of the customer. I think that some of the most important work before us today is to strip away the apparent complexity around what we mean by productivity, including how to measure it, and to put those metrics in alignment from the Pull Request Edge all the way through to the P&L.
Time is our most valuable asset and I thank you so much for devoting a bit yours to my humble account. I hope that you’ve found something in it to be of inspiration in your own journey.
Modern Building Photo credit: Alex Wong
Cafe Street Scene Photo credit: mana5280Tweet alt="Tweet this" />