Thread by @PNJeffries

This thread by @eng_mclare has (deservedly) got a lot of attention and as it chimes with a lot of what we've been discussing in @ramboll_uk about #skills, retention and the need to '#digitalize' in #AEC I thought I'd share some thoughts on how we can do that. THREAD ALERT! 1/

https://twitter.com/eng_mclare/status/1310224676141047808

You can consider this thread a sequel to my talk on 'Computational Design at Scale' at the @IStructE last year: . This is what we need to do at an *even bigger* scale. All opinions my own etc. 2/

Let's take it as-read that the industry *needs* to digitalize, although I'm very aware that there are still quite a few... um... 'more experienced' engineers who might still wish to debate that. They are, however, decreasing in number one way or the other. 3/

To any hold-outs my argument is simply that one of two futures await us. Either we learn to do what the tech industry does... or they will learn to do what we do. I invite you to consider which of those is preferable from an institutional knowledge and safety perspective. 4/

So, how do we digitalize? Step one: stop talking about 'digitalization'! I'm obviously being a hypocrite by using the term in this thread, but I'm doing so with due consideration as a catch-all because what I'm about to say is very general and broadly applicable. 5/

However, 'digital' is a bit of a nothing term nowadays - everything we do is digital whether we like it or not. We need to be more specific. Are we talking computational design, automation, visualisation, digital products, data science, #BIM, sending emails? 6/

There's crossover, but all of these different things have different required skillsets. There's not just one set of 'digital skills' anymore. My experience is that getting a lot of 'digital' people together results in a lot of general agreement but little concrete progress. 7/

This is because the actual interventions required to make progress are many and different. You need to have specific business goals to focus on and work towards, not just a vague commitment to becoming 'more digital'. 8/

I haven't quite succeeded in banning the word 'digital' from @ramboll_uk just yet, but some people have started apologising to me every time they say it, which is progress! 9/

However, beyond the specific skill requirements for each sub-discipline the most important thing is the blend of those skills and the *degrees of specialization*. Which, ideally, looks something a bit like this: DIAGRAM TIME! 10/

At the bottom of this pyramid, you have your non-specialist engineers/architects/whateverers. They need a base level of whatever digital skillset we're talking about (probably much higher than most of them have now), but it isn't their focus. 11/

Worth noting that they will have other specializations and equally valuable skillsets - this is a pyramid of increasing digital specialization, not increasing worth. They win and deliver projects and they keep the lights on. 12/

Up the top, you have your dedicated digital specialists. Your meganerds. Hi there. They dream in code and poop digital workflows. Their job is to keep pushing the tech forwards. 13/

They may not even work on projects at all any more (or mainly work on silly weird pavilions that don't make any money) - it doesn't matter, their job is mainly R&D and to be troubleshooters and a knowledge resource for the rest of the business. 14/

In the middle you have the 'hybrids' - people with higher-than-average digital skills and the ability to apply them to real design problems at scale. They combine domain knowledge and digital kung-fu in a single brain and can spot opportunities and needs for new tech. 15/

They also are key to scaling and realising the value of the top-tier tech-heads, spreading new tech downwards and outwards and business needs upwards. As such they are probably the most important layer. 16/

(Actually, in a truly utopian far future, they might even be the most numerous group and you'd have a diamond rather than a pyramid, but let's learn to walk before we try to run...) 17/

This thread has no gotten a little bit longer than I was anticipating and we still have a long way to go, so I'm going to park it there for the evening and come back and do a 'part 2' tomorrow. 18/

I will leave you with a SHOCKING CLIFFHANGER TWIST REVEAL that the distribution of skills we actually have in the industry doesn't look like this pyramid. It looks, instead, a bit more like this: 19/

Join me tomorrow for some thoughts on why this is, the issues it causes and what we can do about it. Is the industry doomed? Will robots eat us? Will I actually make this somehow relevant to @eng_mclare's thread? Find out then... 20/

AND NOW... THE NEXT EXCITING CHAPTER... 21/

So, yesterday I talked a bit about the skills distribution we need vs the skills distribution we actually have in #AEC. The latter was probably even a little generous - a lot of smaller firms might not have any of the top two tiers at all. 22/

So, how do we bulk up the top two tiers of that pyramid? First, we need to get people to the level that they want to be at, and this is through education, training and support. This is very sub-domain-specific so I won't go into much more detail here. 23/

However, generally speaking, there is a huge amount of educational material already out there (and I'm currently pondering putting some more out there myself in the future); it's pretty easy to self-teach a lot of very useful digital skills. 24/

What is *not* easy, as @eng_mclare touches on, is getting the time to actually do this. Projects, and the AEC professionals working on them, are under enormous time pressure and education and training time gets squeezed out. 25/

Firms need to make more time for training and experimentation on projects, but are *extremely* reluctant to do so, because margins are also shrinking and billing ratios have to be maximised. 26/

Why are margins shrinking? Because, in the words of one property investor I talked to: "We don't really care that much about the structural engineering. We trust you guys to get it right and if you don't, we can just sue you." 27/

In other words: our value-add is not always being felt by the people who ultimately pay for our services, besides as a financial risk management commodity. This means we all end up competing against each other on price, and it becomes a death-spiral race to the bottom. 28/

How do we fix this? Well, we need Land Value and Carbon Taxes to re-balance our economy and industry away from its basis on land hoarding and restore the value of quality building design. 29/

Oh yes, you've just been STEALTH GEORGIST'D!

But assuming our political climate does not suddenly become sensible overnight; what are some other practical steps that are more within our control? 30/

Well, 'digitalisation' is the goal we've set ahead of ourselves, but it's also, partly, the solution. Automation and Computational Design technologies can increase margins by reducing pointless busywork and opening up new routes to providing value to clients. 31/

Also, conveniently, the absolute best way *by far* to develop these skills is through using them on real-world problems. 32/

This means that what is needed is an up-front investment in developing those capabilities but after a while they start to pay for themselves... *provided* you can scale and sustain them through the correct skills distribution detailed above. 33/

As an example; once you reach a certain level of proficiency with Grasshopper, it becomes just as fast to model most things that way rather than manually in Rhino... 34/

...but with the added bonus to the project that the model is parametric and the added bonus to *you* that you're practicing and developing computational design skills. 35/

So, businesses don't *necessarily* need to sustain a high level of investment for long periods, but they do need the *vision* to make that initial investment of time and money in upskilling their people. 36/

That time investment is key to start pushing up the peak of that skills pyramid. However, it is necessary but not sufficient; if that is all you do you'll need to *keep* pumping money in because the reaction is not yet self-sustaining. 37/

The next mistake businesses make is that they do not distribute that investment proportionally. Typically, in AEC, if you're good at something you get asked to do it again, and again, and again, until you are The Person That Is Good At That Thing. 38/

This might get you to one individual at the top of the pyramid, but they'll just be hanging there in space, disconnected from everybody else without the necessary support below them and sooner or later, they're gunna fall down again. 39/

You want that person (might even want to *be* that person), but you also want to spread the opportunities around a bit as well in order to develop the middle tier. This tends not to happen and is one reason that middle band is the smallest in my 'actual case' diagram above. 40/

This will all help you build the structure of your pyramid. But the other thing you have to do is maintain and stabilise that structure, and THIS is actually both the most crucial part of the whole equation *and* the thing that AEC is mostly getting so, so, very wrong. 41/

And so THIS... will be my topic tomorrow night, in Part C of this thread, because I originally thought this would be about 5 tweets long and *clearly* I miscalculated...

So join me then, for some CONTROVERSIAL OPINIONS™ as we solve the riddle of why even though #AEC can sometimes produce digitally-gifted people it is not super-great at hanging on to them...