One of the cool things about IIeX is how it’s always keen to try different things with the typical conference structure. And for this year it shifted the venue to the University of Texas, what better start than a day of education? Five tracks of workshops and masterclasses, covering a huge variety of topics.
The workshop format is a tricky one for any presenter. First of all, you have to balance teaching and interaction. Second, you have to fill an hour, not the usual 20 or 30 minutes. And third, you have to work out where your audience is likely to be in whatever learning journey you want to take them on.
All four workshops I went to had the balance slightly different. But overall they had come to similar conclusions: an hour-long slot is a chance to present some of the basics, in an engaging and fun way. But also to put your own spin on them – offering a strong point of view so that even experts will learn something new.
Here’s what I saw, and what I learned.
Hitting the Bullseye with the Scale-House
The first session was all about growth strategy – Jackie Anderson from ScaleHouse taking us through her three-stage X-R-M process for working out what you need to do and how you need to do it.
X – using a matrix of internal/external and people/products to work out where the barriers holding your growth back are.
R – understanding whether what you need is a recharge (doing the same better), a reboot (starting over), or a redirection (doing things differently).
M – measuring progress and creating accountability with a consistent, company-wide framework of goals
The strength of this presentation was making things simple, but not simplistic. Strategizing for rapid company growth, or for scaling up, is a beast of a job. There are a lot of people – mostly with books to sell – who will suggest there’s some kind of secret sauce which can kickstart growth. The truth is that you need to look to the basic ingredients before you start worrying about adding sauce.
This session put the critical decisions leaders have to make into frameworks which clarified them without downplaying the difficulty. As the final section on goals made clear, you can’t avoid getting to grips with the detail of what you need to do to grow. You can make sure the detail is all in alignment.
Understanding Customers with Little Bird Marketing
The second was Priscilla McKinney of Little Bird Marketing, with a presentation filled with rainbow unicorns and straight talking. (And worksheets! “I love a worksheet”, sighed my neighbor, happily.)
The straight-talking first. B2B content marketers – which means everyone, pretty much, as everything you write, say, or send gives signals about your brand – have to learn the same lessons B2C content marketers do. IT’S NOT ABOUT YOU. Nobody cares about your brand, your story, your methods, your news or what makes you special because, guess what, you’re not special.
But you might become special if you can understand what makes your customer special. The focus of the session was on doing this via building personas – verbal voodoo dolls of your ideal clients. Get to know the personas’ hopes, fears, and needs and you can craft marketing copy which actually feels relevant to their problems.
(Or that’s the idea. From experience, it’s all too easy for confirmation bias to creep in and for your personas to just so happen to “want” exactly what you’re selling. But if you can dodge that temptation, you’ll be fine.)
This was a session which brought that hoary old phrase, customer-centric, to vivid life. It was easily the most interactive of the four, with a lot of fun and helpful exercises and a sense of purpose. Understanding clients better isn’t about pandering to them – you can keep a unique tone and style – it’s just about making yourself heard when everyone around you is shouting.
Making Better Ads with System1
Yes, I work for System1. But I still learned a lot from this session. Because, to be candid, you learn most about your product when you see people use it and what they get out of it. And this was my first chance to see System1 Ad Ratings being used, not just by a client, but by the audience too.
Ad Ratings is a tool for – well, that’s the thing. We built a product knowing full well that we were only predicting what it might be used for – we wouldn’t know until people started doing it. Ad Ratings is a dataset – a continually updated collection of effectiveness data, based on testing every US and UK TV ad that airs. It’s a joy to explore and see real consumers’ verdict on ads for everything from candy bars to suppositories. But it’s only as useful as the questions you bring to it.
That’s why it was so important that we had Logan Moorse of Post Consumer Brands onstage with System1’s Jocelyn Simon. Moorse had a problem and was using Ad Ratings to help solve it. After 40 years as a worker for Post’s Honey Bunches of Oats brand – and then as the star of its commercials – Diana Hunter had retired. Ad Ratings data showed what Post suspected – people loved Diana’s ads, and the challenge for Post was to sustain that level of affection without her.
In this session, System1 opened the cereals part of the database to the crowd – encouraging workshop participants to dig around the good and bad of the cereals category to get a feel for what worked and suggest what Post might try next. Even with just ten minutes of work comparing ads and exploring category codes, the audience came up with some great ideas. For Moorse and Post, the speed of Ad Ratings is a real advantage, but this session showed how powerful the combination of large-scale research data and human creativity can be.
Helping Machines Learn with Converseon
Finally, one of the hottest topics of the whole conference, AI and machine learning – and a workshop from Rob Key and Ben Sigerson of Converseon, who are looking to ‘democratize’ AI.
We’re sixty years into the study of machine understanding of human language, and making advances all the time. Algorithmic deep learning represents a step change from the rules-based language analysis most ‘text analytics’ tools rely on (or used to). Crucially, though, the algorithms get better, faster, when humans have a hand in training them.
This human-assisted learning is what Converseon is working with – “active learning” which can use human input to train more accurate models quickly by asking people to classify both edge cases and those the machine is already sure of. In the workshop, this classification was done by the audience, around brand trust and mistrust in the automotive sector.
A technical hitch meant we couldn’t see which audience member had synced up most accurately with the machine mind– but the session was still fascinating (and the results and winner were announced soon after). Democratizing machine learning to bring it out of the preserve of data scientists is a holy grail for the insights industry.
Converseon’s main insight isn’t technological, it’s about business models. The problem isn’t that the algorithms are complex, it’s that setting up an accurate model is long, technical and tedious. Converseon offers the algorithmic equivalent of fast food – a menu of pre-trained models which can then be let loose on a specific dataset and adapted to it with human help.
Overall, a very strong opening to IIeX – a lot of learning condensed into four 1-hour sessions, and a very civilized 11 AM start. With a few administrative tweaks – making it clear what devices you should bring for each workshop, for instance – “workshops day” could easily be a true annual IIeX highlight.