Turing Festival, 2019, Edinburgh, UK

A fixture of the Edinburgh tech calendar, I look forward to Turing Fest every year.

This year’s edition had 3 tracks: Build, Grow, Lead.

As a marketer, I was obviously mostly drawn to the ‘grow’ track, but it’s important to expand your horizons so I made sure to listen to some topics outside of my usual focus.

It’s good to get some fresh perspectives and new ways of thinking.

Here are the highlights of my Turing Fest 2019:

What Happens When Every Website is Fixed? Jono Alderson.

Jono Alderson described a future where everyone’s technical SEO issues were all fixed. What happens then?

He suggested that quality is becoming quantifiable. Google is getting better and better at determining what context is actually ‘good’.

Because of this, Google is taking over. People trust its search results. They get the information they need.

To make things even easier for its searchers, Google is starting to show zero click search results: where information is displayed right on the search results window.

From Google’s perspective, is your website friction? Is it stopping users getting the information they need as quickly as possible?

If you want to show up in any search result, let alone the on-page search result, you must understand the importance of structured data.

Google is a huge proponent of standard setting on the schema.org project. These are essentially the ‘rules of data’, where each piece of information should be tagged properly so Google knows what it is.

Google and WordPress have teamed up. WordPress powers a huge proportion of websites on the internet, and the two organisations are making structured data protocols standard within WordPress. That’s part of what WordPress’s block editor, Gutenberg, is all about. See also: Sitekit.

Alderson emphasised:

• The importance of site speed – so Google can crawl faster.
• Progressive web apps – so Google can see inside your app

Ultimately, when your data is structured, you can focus on trust and fit rather than the technical stuff. This means you can spend your time giving customers/searchers what they need, rather spend time and energy on technical SEO.

Is this a dystopian future where Google owns the internet? Maybe. But what are your alternatives?! 

A slightly frightening but very enlightening talk!

Slide deck available here.

Hack the Stack: Five Ways to Automate Like a Human. Aiden Carroll.

Aiden Carroll gave his talk on automation structured around five themes:

Ambition

Audience

Message

Format

Measurement

Ambition:

As marketers know, you must strive to deliver the right message at the right time to the right audience.

There is currently a mismatch between ‘branding’ and ‘programmatic’. Branding is too often irrelevant to your customer and mostly about you. Programmatic campaigns seem creepy as they analyze your shopping basket and follow you around the internet.

The balance is: Don’t annoy people but, at the same time, don’t annoy your boss. You need to find the balance within the overlap of meeting human needs – being useful and relevant – and meeting your business needs – meeting performance targets and business objectives.

Audience:

To state the obvious: Target the right people.

Remarketing data allows you to reach people based on prior behaviour at your website.

Media platforms allow you to reach people based on their affinities: things they like.

Message 

You need to recognise that different personas have different questions and different needs.

You cannot assume that the same marketing will be relevant to different personas: Be clear who you are talking to.

For example, if you were marketing a music app, your approach would be very for a tech geek as opposed to a concert lover. One would want the technical aspects emphasised, the other the quality of the sound and the concert-like experience.

Format

Not all ad formats are equal. Based on the message and the audience, you should choose the best ad experience. Aiden recommends skippable video and non-intrusive display ads as a starting point.

With your ads, you need to find the overlap between helpful and impactful. Don’t be noise!

Measurement

Any campaign requires clear KPIs:

• Did I get a good price?
• Was it shown to who I wanted?
• Did people actually see/watch it?
• Did people engage with it?
• Did I come out on top?

When you set KPIs, you should analyse them at the end of a campaign. Did you hit your goals? Why or why not?

Conclusion: Unite your geeks, suits, and creatives! A successful team needs all of these perspectives to succeed. Alignment is hard, but you need all these people working together in order to get results.

Slide deck available here.

Three Growth Choices that Can Scale Or Kill Your Business. Kieran Flanagan.

Kieran Flanagan spoke on how to achieve growth via marketing. He stressed that growth is a roller coaster; success comes in fits and starts, and after experimentation.

Growth is much more than hacks or tactics. You need to be clear on 3 areas: 

• Language
• Team
• Strengths

Language

Your growth teams must align around a common terminology.

When you bring a group of people together, they need to share language on how they intend to work together and what they intend to work towards.

Kieran suggests you get the team to create a service level agreement (SLA). This means that a work standard is agreed up front, which reduces misunderstandings and miscommunication.

Everyone on the team needs to agree on the problem and agree on the goals. A shared language makes that much easier.

Team 

The growth team needs to align around specific opportunities for growth. They form to tackle a problem and dissolve when it’s complete.

In Kieran’s experience, at a certain size team structure is the limiting factor for growth. The organisational structure gets too complicated and unwieldy for quick victories. Instead, cross-functional teams should be created to focus on growth ideas (Including Dev, UX, PM, SEO, etc)

Achieving alignment between all these disparate is HARD (hence the language bit above!).

It’s better to overinvest in a few promising opportunities, rather than spread resources across everything. You get little done when you spread yourself too thin.

Strengths

When prioritising which growth teams should form, lean heavily into your team’s strengths. It makes no sense to take on needlessly hard projects. Do what you do best.

Kieran gave the example of a HubSpot growth project. The team was strong on content, storytelling, and SEO. So the team focused on creating content to win ‘hearts and minds’ for Hubspot. Guess what? It was a success.

Something to note: You do need to invest in your team’s learning so that you are able to create future strengths to leverage in future growth projects.

One final point: Kieran noted an ‘internal startup funding model’. Small teams get resources to get their growth projects started. But they then need to justify further investment but showing significant progress. This seemed like a good way to balance the need for experimentation with allocating scarce resources.

Slide deck available here.

Machine Learning for Marketers. Britney Muller.

Britney Muller gave an update on how rapid advances in machine learning might be leveraged by marketers. Because machine learning is becoming mainstream, it’s becoming more accessible.

There’s a lot of scepticism about how AI will influence the world of work, but Britany’s view is clear: Machine learning frees up time/effort for higher level strategy. The more ‘grunt work’ we can let computers take care of, the more time we have to work on more important things.

What exactly is machine learning? Machine learning is when a programme can complete tasks by ‘learning without being programmed’.

You must provide a good dataset for the algorithm to be effective. As the saying goes: garbage in, garbage out. Not only is the quality of data important, so are the biases it might contain. We don’t want machines to learn our own prejudices, and this is a real problem, currently.

If you can overcome these important and difficult issues, the good news is that you don’t need to be a data scientist to take advantage of the power of machine learning.

Britney gave some real world examples you can explore today:

• Frase.io – use existing content to power live chat
• Lumen5 – automate videos
• Amazon Transcribe – automate transcription
• Numerous natural language algorithms – pop-the-hood and see how Google (and others) classifies your text

Britney gave an excellent state-of-the-field summary. It’ll be exciting to see how many tasks machine learning powered tools can take on for marketers in the years to come. 

Check out Britany’s slides for even more resources.

Death by Design: Killing your Growth with Defective Device Experiences. Craig Sullivan.

Craig Sullivan’s talk was perhaps my favourite of the whole conference. His key idea was clear and simple: Bad design kills growth.

When it comes to bad design and user experience, people don’t complain – they abandon your website/app and disappear forever.

Machine learning isn’t good enough yet (see previous talk by Britney Muller!). Craig said that machine learning in its current state leads to bad experiences and, put simply, ‘scales up stupidity’.

Craif introduced a new term, a ‘UX bug’ as opposed to a software bug. What is a UX bug?:

• Can’t scroll
• Can’t press button
• Text too big for the screen
• Etc.

There are often very simple fixes to these issues. For example, many terrible forms could be fixed by changing the touch keyboard settings: 

• Disable autocorrect, 
• Use numeric keypad for phone numbers 
• Etc.

These simple fixes on a website with high could be huge!

As with bug testing on code, you need a ‘bug hunter’ approach to UX.

How do you go about deciding where to look for these UX bugs? Look at user analytics: Use the most used devices to prioritise testing.  

• What devices are most frequently used by you visitors? 
• What device / screen combination had above/below conversion rates? Why?
• Etc.

For more testing ideas, see: Rapid mining for eCommerce Conversion

Quick hits on the UX bug squasher approach:

• Use your OWN stats to drive testing.
• Draw up your OWN testing list
• Test regularly, especially after any updates
• A/B tests, if thought they’re experimental, should still pass QA

Final thoughts: 

Accessibility expands your reach: Don’t optimise for rich, 50 year old iPhone/Mac users!

Growth is a strategy not a result

Slide deck available here.