People seem to get really carried away with (or perhaps obsessed with) the most unusual tiny aspects of producing espresso. These can range from the brew temperature of the espresso machine (it's important, but do you really think that the problem with the shots you're pulling is that you don't have electronic temperature control on your machine?) to arcane and complicated tamping techniques to bizarre cleaning rituals and beyond.
The reality is that there are more than 100 variables going into a single espresso. As a result, it's all a dance. Or, perhaps more accurately, it's one of those plate-spinning acts while dancing with a partner. Hmm... yeah, pretty complex.
Anyway... we try to be consistent in controlling as many variables as is humanly possible. But there is always variance and there is always going to be variance. So... I think the key is to pay attention to the most critical variables, and then understand the relationships between all of the variables and how they affect the results. Understanding this allows you to then pay attention to the results in light of all of this (the key variables and the inter-relationships between all the variables and their outcomes).
So what are the key variables? Well, there are some that can be reduced to constants like brew temperature and brew pressure and the coffee and the consistency of grind. So let's assume you're working with acceptable equipment and high-quality, fresh coffee. Given this, what are the key variables for a barista to focus on?
Grind is the one variable that should remain a variable. The goal in general is to try to be so consistent and controlled at a tiny detail level that all other variables approach being constant. Now... of course... this is very Zen. I mean, it's basically impossible (and thus the idea of understanding the relationships between all the variables). In any event, grind is the variable you use. By this I mean that you tune the extraction of your espresso by adjusting the grind. Learning how changes in grind affect things, striving to understand the complexities of the grind is 90% of being a barista.
Distribution is the most misunderstood, neglected and really critical variable within your control. For good espresso, a requirement and the goal is to create an even density of coffee within the basket. This is actually a non-trivial problem. For most baristas (professional or passionate enthusiast) this is where technique fails most noticably. There are a couple of options when it comes to methodologies for distributing coffee within the basket. The two that seem the easiest to grasp, most predictable and most reliable are the Schomer Method and the Stockfleths Move (video). The best thing to do is figure out which one works for you, practice it, and get good at it. Keep in mind all the time what the goal is... to create an even bed (an even density of coffee within the basket). If you get this right, then your odds of correct extraction are going to go way, way up.
For each coffee, there is a range for your dosage that works with a set combination of grind, distribution, tamp and extraction style. You simply have to stay within that window. For target tolerance, 0.1 grams is desirable but probably excessive, but 0.5grams is really the maximum variance that will work. It's worthwhile to actually practice your dosing with an eye towards this tolerance. A good thing to try is to grind, dose and distribute your coffee in a clean and totally dry portafilter, and then dump it into a clean, dry paper cup. Repeat nine times. Now weigh each one. Practice until your tolerance gets to a personal target (I'd suggest 0.3 grams).
To be honest, everything else is less important when it comes to a constant evaluation. For example, the tamp is not nearly as important as everyone seems to think. The goal with your tamping is to preserve the distribution and to create a firm enough surface. You need to be consistent with your tamp, but then again you need to be consistent with everything you do.
Preservation of distribution is an important thing to keep in mind. There is no point in creating a near-even distribution only to then destroy it. How can you destroy good distribution? The most common ways are: whacking the hell out of the portafilter (we've all seen baristas who seem to be performing some percussion piece using the tamper on the portafilter); tamping off-level; rattling the portafilter side to side in the grinder fork; tamping multiple times with a too-small tamper; etc.
Anyway... back to tamping. Basically, if you preserve the distribution, tamp at least 30lbs of pressure, tamp with consistent pressure and don't mangle the coffee by polishing under pressure, tilting the surface, etc., it's all good.
So... other "obsessed upon" variables? Well, time and volume are then going to be tools to help you get close to the right grind. When you get close — you start tasting while tweaking the grind to get yourself to the point where the grind is right. At that point, however, throw away the tools!! This is an art as well as a science and is meant to be tasted as a method of evaluation. So instead of watching the stopwatch, watch the flow of espresso. Even when you're being as consistent as you can possibly be, there will be variances (not just in your technique but in other factors). So in watching the flow, you manage volume and time based on that flow. And taste, taste and taste some more. Because at the end of the day... it's the taste of what's in the cup (not the size, not the time, not the look, blah blah blah) that matters.
Drink the coffee — love the taste.