CTRL-Shift-Escape - #6

What Are We Trying To Do?

In the last year or so we’ve been restructuring a bit, creating some new teams, some new management positions and generally making room for continuing to grow. Kate Raynor pointed out to me recently that although there’s a generally good ‘buzz’ in the whole company, we need everyone to know what the plan is. With going on for 60 people in the business, many of them remote, we can’t expect it to just percolate through the company like it used to as we chat around the water cooler (we don't even have a water cooler).

“Well that’s easy to fix” was my first thought. We’ve got a business plan for 3 years, which we review regularly. That’s what we are trying to do. It’s a really good business plan (I may be slightly biased) and it’s very exciting. To me.

It’s probably less exciting to someone who is working with a frustrated client trying to debug a weird issue with a missed bin process. When the guys are working a weekend in 30 degrees heat fitting telematics to 30 smelly bin trucks, it may not be the greatest motivator.

So, just like you need an ‘elevator pitch’ in sales, we need one for the business. When I’m making the coffee and someone says “How’s it going?” I need to be able to tell them, and in terms that join up my business plan to their working day. And when they’ve got a problem that they need my help with, we both need to be able to see it in the context of how it affects us all.

OKR

This is where Objectives and Key Results comes in. It’s a framework for setting big goals, and then cascading them throughout the business. Like most of these things, it’s common sense written down but it’s my big focus for the coming week.

The basic principle is that you set a general 'objective' which is aspirational and qualitative. Maybe "Launch a new product that does X". You then set out a handful of specific measures with a timeline that will mark that objective, or at least your progress towards it.

We will get a handful of them written down at the high level, and then work with the other teams to break them down in smaller ones that can be shared across the company. If we get it right then we’ll have a set of objectives that span across our teams rather than creating silos of people with competing incentives.

Mailing Lists - A Major Fail

This week’s adventure in HubSpot was the first publish of our newsletter through the platform. Our newly cleansed contact list, finally merging all our various databases and spreadsheets was ready to go. Nina Ramskill had the copy ready to go, and so we hit ‘Send’ and ticked another job of the list.

After 250 sends HubSpot spat the dummy and said “Nope”. 20% hard bounces (we normally a handful in total). So something has gone wrong with the logic, with the data or with the consent / subscription data. Maybe all three!

We’ve ‘unticked’ the ‘Send Newsletter’ item on the to do list, and added a new one - “WTF has happened here?”. Fortunately we haven’t turned off Mailchimp yet, so we can revert to that and try again next month….

If you want to subscribe to our newsletter - well, actually, I'm not quite sure how that works at the moment!

Hackathon

Until this week I’d never been to a “hackathon”. From a distance they looked like quite intimidating events where rockstars coders with zero social skills would collide with management gurus to cover a wall with post-it notes, drink lots of vending machine coffee and write a blog post.

Having spent 2 days with Birmingham City Council and Amazon Web Services at their #FoundryLive event this week I can say that I was mostly wrong. The rock star coders were there, so were the post-it notes and the coffee was average at best. And I’ve written this blog post about (and so have many others). The pleasant surprise was the enthusiasm, the sharing and communications. We didn’t win a prize, but we did complete a genuinely useful proof of concept. I genuinely think it could save lives.

The team created a process which uses AI to process video footage from the back of a bin truck. It’s trained to spot two common safety problems - loaders entering the ‘crush zone’ (where they are at high risk of being lifted into the compactor) and crossing the road on a ‘single sided’ street (where they are at high risk of being run over). Both have led to fatal injuries in recent years.

It’s impossible to watch everyone working all the time. But AI can. We fed it a random hour or so of pre-recorded video and it spotted 20 (yes, two zero) safety issues. Extrapolate that to over 10,000 bin trucks working on British streets every day for 6-7 hours and there are huge gains to be made. Currently we do this manually (and with great outcomes) - with this technology we can increase the effectiveness by 2 orders of magnitude.

It’s a proof of concept, and there’s a lot to do to make it really workable (and affordable).