Modern CTO Podcast
Guest: Andrew Ward
Today we are talking to Andrew Ward CTO of modl. We talk about What Makes Great Leaders, Machine Learning, Asking for Feedback and much more. This episode is high energy and packed full of knowledge. All right here right now on the Modern CTO Podcast.
Time Stamped Notes
1:03 Machine Learning
When it comes to Machine Learning, building it into a thing and not knowing what to do with it has been what has been a repeat practice amongst companies. In fact, “Machine Learning,” has simply become a buzz word with innovation as the assumption is that it makes their starter more valuable. Andrew encourages companies to stop thinking about Machine Learning and to start thinking about predictions. What is machine learning without predictions you’re trying to test? A non tech person approaching the machine learning topic just thinks that a computer will do it for them. There is no distinction between what’s AI, what’s an automated system and what are the types of machine learning. There’s a lot of capabilities and benefits. They’re aren’t that many machine learning specialists in the UK at least. With Large Scale Organizations, there is a tendency to look at what data they have at the moment and how you can make machines based upon that data. How many unsolved problems are there to solve the machine learning problems and are readily accessible? The probably may not be the training of the computer – but you need that meaningful data to solve the problem. The play that’s least discussed is having high volumes of quality curated data for the models to train themselves on. If you want to teach machine learning about empathy, you’ll need 100’s of varying stories where empathy is the centre of the story. Then that can be consumed by a machine learning model. You need quality data based on what you want it to do so you know that it was made for that purpose.
9:17 Base Programming Language
10:38 Code Structure Examples
Historically what’s happened with programming languages is if you go back ten years in the php world – it used to be a very useful lose typed language. It was considered the type of language used when you are hacking away in your back shed. You have flexibility to do what you like with it which meant you could write very bad code very easily. When Java came out – it was a typed language and you had to be strict on certain conventions you use and certain concepts were popularized, especially coming out of university with the Java. This is where Java lead the standards behind programming language that were considered to be not very good like php.
11:23 Framework of Choice
Andrew’s framework of choice is between E2 or Symphony. Joel talks about how he found his favorite framework in Ruby. Andrew stayed on the php train as he believes when you are growing a company from being just him, he didn’t really have the credibility to take on large projects. At the time, his company were building the team around web based solutions – a lot of the early day stuff was pumping out websites. That’s where the php journey started but nowadays with the tools that exist for it, there are platforms that exist to support the hosting scale well without technology.
13:12 Heroku & React
Joel talks about how he has some apps on Heroku. Andrew explains his experience with Heroku and talks about how much he loves React. He runs some projects on Angular 1 and 2 however finds React has been up there as a framework everybody loves using. Joel talks about how he tested his apps on React Native. It’s made his life easier.
15:38 What’s Right Versus What’s Easy
Joel took a long time to get into React with both web component and react native projects. He was burnt by previous garbage frameworks which caused infinite stress. From a marketing perspective, it was hard for him to be sold on it and then saw developers using it. After trying it on an internal small project, Joel was blown away. He then tried this on a new client. He was open and honest with the client. Testing was done on an internal project. This was also done with a price reduction and it was the communication that allowed it to happened. Andrew explains his experience around React too. When he first saw it he thought “Oh, this looks easy.” But then he had to stop himself to see if it was the right solution.
18:17 – Scorchsoft & MODL
Through Scorchsoft, it allowed Andrew to be in the tech scene for the past 7 years. It helped him build an ecosystem formed in the Edinburgh area. There were some friends of his, also entrepreneurs who began talking about a problem that existed in the modeling industry. There was an outdated marketplace, high fragmentation and small rosters of models. From a models perspective – its not uncommon to receive payment between 30-60 days for a job if they receive it at all. Everything is email based. It’s really not a modern up to date process. Andrew and his friends sat down and found that this was a meaningful problem to solve. The team were well experienced and had great skill sets to bring to the table and it was found to be where an alignment of skills came together.
23:53 Communication & Technology
It’s quite a skill to be able to communicate technology in a way that is understood by people of different intellect and artistic abilities. From a visionary side perspective – these particular people can be any where on the spectrum. You have to be able to communicate to them pretty advanced and complex topics in the way that they understand and make decisions on the basis of. It’s not just the case of knowing the tech, but knowing the person and reading the situation. That’s a way to be an effective tech leader.
31:53 What Makes Great Leaders
In terms of what makes an effective senior manager – you look at the CTO’s, CEO’s and CMO’s – they are never down in the trenches doing the doing. And they shouldn’t be. They are communicating and strategizing and managing people. Joel discusses his own journey to being a CTO. He explains how he found it difficult to qualify or quantify the work. Previously, he had done it through progression throughout the code base. That sense of completeness as a developer found him scoffing at the business people or the business management side of things. It was when he got into the business side of things that he realized – it is real work. It’s a lot of effort and energy to come up with the simple communication and relay the same information with the same consistency and energy. It’s a whole different psychology to how you motivate and drive yourself to what tasks are meaningful. When you’re in business, you’ve got to constantly evaluate if what you’re doing is worth the time. When you’re transferring from a developer to a business leader, it’s almost a comfort to stay in development.
34:55 Constraints around Delegation
From a perspective of an agency model, if you’re a lone ranger in programming, you become very good at freelance contracting roles. This is where you’re in a position where you are touching on the whole entrepreneurship game, but only touching on it as they are essentially working for somebody.
What Andrew has seen with agencies who want to grow is they take on more tech people in the business and become a team of 2 or 3 and they begin to take on larger tech projects. If you want to step away from being a developer in that business – that step is going from 3 developers to 3 developers and a 100% fee burning head. The business has to change to support that overhead. You may have to review your existing rates or you have to have a change in the way you do business. It doesn’t stop there. That’s the first time you’ve had to deal with a fee earner to a fee burner. If you want to grow you may need an account manager or project manager, unless you have a solid structure to support that growth.
The lone rangers are completed booked solid and can’t expand.
There is also a constraint with developers that become CTO’s – the big problem that keeps coming up is that they are not able to step away from the development. But then they have issues either letting go because they quantify their value. The roles and responsibilities of a business manager is so different to being a technical engineer. There is an inherent misalignment there. It’s hard to transition to understanding what motivates people. People who code love to solve problems and get their hands dirty and love to test and play with what they have built. As a CTO, it’s difficult to know what you need to focus on to stay in the game and what you need to step away from. As a CTO you’re always problem solving – it’s a foundation of a computer science that you need to know to solve that tech problem.
Pair programming with developers as a means of making sure their constantly problem solving in the code so he’s not in too deep.
46:17 Evaluation of Competence
There are four types of awareness people tend to fall in when it comes to awareness of their own competencies and skill sets:
Unconsciously Incompetent: People who are unaware that they are incompetent. This usually means that present a lot of bravado and ego around their work however their results and deliverables do not meet the requirements.
Consciously Incompetent: People who are aware that they do not have the skill set to deliver a task
Unconsciously Competent: People who are not aware that they possess skills and attributes. This generally means they are doing things on auto pilot.
Consciously Competent: People who are aware of their own strengths and skill sets. These people generally really thinking about what they do.
This applies to tech because people generally are very good at what they do but given the amount of technology out there, everyone may use different language. People don’t see what they know, but feel anxiety when they don’t know. It makes experts feel incompetent. They don’t realise the value of the knowledge they have.
48:15 Problem Solving
Problem solving is a negotiation. When you’re in a negotiation with someone, you are trying to get to engage their pre frontal cortex. When you’re problem solving, if someone has got anxiety against the problem, they are not going to solve it very well. Well positioned what and how questions force people to innovate – a series of well-placed what and how is a well-placed tool in problem solving.
Joel takes a very scientific approach to it which a trial and error variation selection. He going tries 10 things, finds the top 3 that work the best and this is also a great when you are trying to figure out what the gaps are.
It’s not questions, it’s recurring framework of trying a handful of things, testing them, seeing which ones work and creating variations of those and find a solution you’re satisfied with. Usually the first solution is not the best. There is a balance. When it comes to the macro, they are going to be the core parts of your business – it’s the approach. People are so eager to launch the perfect things. They are unaware these things cannot be launched on assumptions.
53:43 Signing NDA’s
When people come to Andrew, he signs an NDA because that is what they expect. Even if you have an NDA in place, the reality is, the cost of actually taking out a lawsuit on someone is so high, and the friction it puts into the process of having people have to sign it makes it not worth it. Most people don’t have the time, energy and capability to steal your idea and execute it. However, if it’s an engineering concept or a patent, then it’s necessary. It’s subjective. Andrew wouldn’t let an NDA get in the way of him progressing his ideas.
55:28 Asking for Feedback
Typically, when you call for something analytic in feedback, you put them in a particular state of mind. To be able to get true feedback, ask others what they love about it and you let them start talking to you about it and what will happen naturally is that they will begin to list the things they like and then that one thing they say at the end will have some weight in it’s feedback. This will be a natural negative that will seep up from it. This is generally feedback on a point of improvement. Joel talks about showing apps to friends for feedback. Andrew talks about his model business and how he tested the in person process part of the system. In MODL, what happens is someone puts a job brief through and a lot of models get notified. What then happens on the day is, the model goes to the marketing agency that’s booked them and it’s very likely to have a photo shoot. Some of the first bookings put through were marketing shoots for his own business. Real brands were on board to book real models however Andrew’s team were fulfilling the photography side of things. It was a great way to get good organic feedback.