I'm just back from my fifth Superweek, and this time once again we had a great contingent from Data Runs Deep along for the ride.
We gave ourselves a few days in Budapest to get over the long (very long) flight before heading to the conference, a really wonderful opportunity to spend some time in a city that just gets better every successive day I get to spend there.
Of course, the main event is the conference itself. Up on top of a mountain, with five days of talks and a couple of hundred analyticsy folk to keep things entertaining.
This year, we were at the conference from Monday to Thursday, and the talks were fantastic throughout. With no main focus like GDPR to steal everyone's attention (as seemed to be the case last year) this year's talks were more varied, and somehow more thoughtful.
Monday started with Doug Hall's opening keynote, The Joy Of Data V2. Doug spoke about some interesting case studies and techniques like using GA's client ID to break up your audience into 10% buckets, and scheduling Google Cloud Platform jobs and using a script to alert you when the previous day's BigQuery data becomes available.
Next up was Aurélie Pols with The GDPR Is There So Do You Know What The Courts Are Saying? Aurelie made me think that the industry is occupying a sort of twilight in between GDPR and ePrivacy and that the kind of problems we've been facing will need to be faced again quite soon.
Next I had an unscheduled coughing fit, a legacy from a cold picked up the week I'd spent in the UK, so I took a couple of hours downtime to emerge for Erik Driessen's Lean Measurement, one of my favourite talks of the week. He spoke about structuring analytics projects around a Minimum Viable Measurement Product, usually Pageviews, around which additional custom tracking is added. I also liked his tales about embracing failure and reviewing it on a quarterly basis (something that I expect we could all benefit from doing), and what he had to say about Ans, an automated Slackbot that sounds like an indispensible member of the team.
The rest of Monday I spent working on my talk for Tuesday and, of course, enjoying the catch ups and discussions around and about the place.
Tuesday began with Matt Gershoff's Surprise! It's Entropy, The Theory Of Information. I wanted to describe this as sort of mathematics by stealth, sneaking in some very high-complexity stuff when nobody was looking. Every year I try harder to understand Matt's talks, and every year I suspect he's meanwhile trying to dumb things down for people like me. Either way, you don't need to understand every word of someone's presentation to enjoy it!
Next up was Tim Wilson with Digital Analytics Meets Data Science: Use Cases For Google Analytics. Tim is a fantastic speaker and an extremely personable guy (the two, of course, go hand in hand). Tim's talk was centred around some really smart stuff in R, including mining site search terms that are structured like questions ("how do I..." etc) to get an understanding of what people are really trying to do on your website.
I didn't take many notes through the next talks, as mine was on after -- and that's not to sound like it's not worth paying attention to what someone's saying when I have something to say instead. Quite the opposite!
Speaking at Superweek takes you through so many emotional states, going from relaxed ("these people are all awesome!") to panicky ("I don't know a sodding thing!") to the abject terror that the people speaking right before you will say everything you were going to say, better than you were going to say it, rendering your talk utterly pointless, and probably your entire life as well.
Emotional collapse more or less survived, I returned for Brian Clifton's The State Of Google Analytics Data. Brian spoke about the findings from the various audits he's done on websites, many as part of his new auditing tool Verified Data, with the frankly terrifying statistic that 1 in every 5 sites audited contained PII of some kind. Not cool.
To Wednesday, where the day was very much dominated by Google Tag Manager. Scott Herman and Brian Kuhn from Google held a session that was really interesting, even for someone such as myself that doesn't spend much time at all with GTM. The first thing that really struck me was how gosh darn usable the whole platform is; when you take a step back and look at things like workspaces, approval queues, sharing, and Zones, it's difficult to imagine how anything ever got tracked at all in the days before this was available. It was also great to see the thinking behind, and the execution of, the new Templates feature.
I enjoyed Hussain Mehmood's Flexible Interfaces Doing Data Science Without Code but again I had to duck out as that persistent dry cough was threatening. I can only assume that my cough was being made worse and more lingering by the combination of cold air and heating, and possibly the beer that everyone was making me consume (ahem.) I will however draw your attention to Hussain's tool utm.io that appears set for a very tempting and very exciting reboot.
The weather was stunning on Wednesday, so we went out wandering and returned for Simo Ahava with You Can't Spell Measure Without Customization. It was, of course, brilliant and the vibe of expectation and receptiveness was something quite unique at Superweek. Simo was so good he could almost do this for a living :)
Next up was the Golden Punchard, always my favourite bit of Superweek. This year the entries were, again, brilliant with ideas ranging from the helpful (Propensity Modelling in BQML) to the passionate (Avicii's music run through a sentiment analysis and put onto a T-shirt). The winner was Zorin from Escape Studios with a slice of apparent genuis, recreating the multi-datasource capabilities of Google Attribution without needing to wait for the release of the tool.
Wednesday night is usually the biggest night of the lot, easing from Punchard to Bonfire by way of Dan Waisberg's birthday bash. Dan announced his retirement from Google Analytics, and while it's a shame to think that I'll be seeing less of a thoroughly nice bloke, it does sound like a really big opportunity.
I decide to celebrate the biggest night of the week by sleeping in far too late the next day, so I miss several talks on Thursday. I also become awash with the awareness that the 17:00 departure time is approaching, and probably spend more time on Thursday hanging around in the bar area talking to people (mostly about analytics, dogs, and music.)
One talk I did attend and found fantastic was Charles Farina's Integrated Marketing With The Google Marketing Platform which spent some time discussing the intersection between Google Marketing Platform and Google Cloud Platform.
This is a gif of Charles. He pronounces "gif" with a soft G:
One thing he closed on really hit home for me -- he said that it's no longer enough to be just a Google Analytics specialist. That stuff has had its day, basically -- and it's no longer cool to just be good at using an interface. We need to learn SQL, play with BigQueryML, and generally take things out of their current safe, stable orbits.
And it's that point that sort of sums up the whole week for me. As I said, there was no "big theme", no GDPR Megalodon from the depths, that we all had to figure out how to content with. In the absence of that, when we don't have to resist any external forces on our work, we can level-up, or move to a new orbit, with the stuff we're doing.
That might mean getting better at Cloud, or it might mean getting better at adding business value to web analytics. Probably, it means both in equal measures.
With so many smart, open and warm people in this industry, I have no doubt that we'll all continue to take it to interesting places.
Events like Superweek are what catalyses everyone and makes stuff like that possible -- and if you ask me why it's my favourite week of the year, that's why.