Managing Marketing: Data, AI And Agile Business Practices

Christian_Bartens

Christian Bartens is possibly best known as the Founder and CEO of Datalicious, but he is also an Investor, Advisor and a C-Level Data Analytics Expert and Entrepreneur who is using his 20 years of experience to help companies build data capabilities in-house that deliver tangible business outcomes. Here he talks about the use of data in marketing, advertising and sales and new opportunities to use data to inform and align businesses to performance.

You can listen to the podcast here:

Follow Managing Marketing on SoundcloudTuneInStitcher, Spotify and Apple Podcast.

Transcription:

Darren:

Welcome to Managing Marketing, a weekly podcast where we sit down and have discussions with thought leaders and innovators in marketing and advertising and this week I get to sit down and have a chat with Christian Bartens. For those who remember, he was the founder and CEO of a company called Datalicious, which not that long ago was bought. First of all, welcome, Christian.

Christian:

Thank you, Darren.

Darren:

It wasn’t that long ago was it that Datalicious was making the rounds of Australia and then you expanded up into Asia as well.

Christian:

That’s right. The company started all the way back in 2007. But in 2014 we attracted some investment from Veda Advantage that helped us grow into Asia and opened offices in New Zealand, Korea, India and Singapore, plenty of spaces.

I ended up going to the US after Veda was acquired by Equifax themselves, expanding the business over there. Those days are over. I’m now out of Datalicious, ready for a new challenge.

Darren:

It’s a great name. Where did you come up with Datalicious? I love it because it says ‘data’ but the ‘licious’ adds this human touch to it.

Christian:

A lot of people think it must have been like a drunken moment in a bar or something like that. But there was a lot of brand research behind it. If I still remember right, we used the MMC brand funnel for that to narrow down a couple of T terms that we wanted to live with.

But one of the key things back then was it needs to be about data but it can’t be dry because it’s bloody boring and nobody wants to deal with that subject. It needs to be a little bit more sexy and that’s how we came up with that. I think back then the General Manager of Hitwise was throwing out these stickers that said ‘datalicious’ on them but nobody had registered the name and boom that was it.

Darren:

That’s the name. It’s interesting you say that people see data as quite dry.

Christian:

I don’t think that’s the case anymore. I think it was back then but now it’s a lot more sexy.

Darren:

Or is it just the new black?

Christian:

Data is sexy, data is the new black—I don’t know. I think it’s pretty mainstream now. I think all of those concepts are passé. Everybody knows they need it. It’s integrated into everything. I think that’s why the Datalicious brand has lost a little bit of its appeal over the years.

We tried very hard to make a bit more of a sea level type brand rather than the initial idea of it being sexy and attractive.

Darren:

We had the discussion about big data and now all the discussions around data are always linked to AI and machine learning and the like, that there’s now this significant shift in the whole conversation around not just data itself but the application of data in things like decision making and artificial intelligence.

Christian:

Oh yeah, you’re totally right. There’s always the next buzzword in any type of industry, right? I didn’t mean to say that data is completely dried out, there’s no innovation happening anymore. I guess what I was trying to say is you don’t have to fight for the budgets anymore that hard.

People are always thinking about data; everybody knows they need to include it. And over the years the focus had changed a little. Like you say, now it’s about AI and automating but in the end that’s just another buzz word.

I don’t think many people are actually using AI and even if you look at some of the software vendors that write AI into the name, there isn’t really any AI underneath; it’s just basic decisioning often.

Darren:

Machine learning; basic algorithms around data-informed but then they learn because it depends on the response that comes back.

Christian:

True. We see a lot of that in advertising, obviously. But, I personally am keen to move out of the advertising space and start using the same kind of technologies and learnings in a different space.

Not that I think advertising is dead; it’s just something I’ve done for a very, very long time and I’m personally more interested in starting to use the same sort of skills in the data space in making teams more efficient.

I think that’s largely born out of my frustrations at the time of starting, building, running Datalicious. Looking back you often get a different perspective and a lot of the things we did were very ineffective. We really didn’t know how the business was running.

That’s something I would like to address. Even though we’re talking about AI, robotics and every human is going to be replaced very soon by machine, I think there is plenty of scope for making interactions in a team smarter with technology and analytics.

Darren:

When people talk about artificial intelligence and they talk about marketing or advertising-related data it’s all about understanding the customer better, personalisation.

It’s about being able to better target, engage, and interact with customers and yet it’s also the thing that’s driven the most concern amongst consumers and also politicians: the very security of that data.

People have started to look at the way it’s applied; first of all the way it’s collected and then the way it’s applied and being used that people think it has some sort of nefarious intention to control and manipulate people. It’s interesting isn’t it?

Christian:

For me, they’re kind of two different issues you bring up there. They are related so let me see if I can explain my thought process here properly.

It’s right that we’re trying to use smart algorithms because we have too much data that we’re not really using. So, when we talk about AI or machine learning we’re handing power of the analysis over to a computer.

There are so many different data points involved that an analyst can’t effectively explore that anymore and turn that into action. What we’re really trying to do is tailor-make advertising for every individual; the right message for the right person at the right time and totally customised and personalised.

At that level of granularity you need a lot of different signals to come up with that right message at the right time. So there’s a lot of AI happening and then you also need to put the right ad together for that particular person, the right message.

So a lot of innovation happening in ad space—building ads automatically based on AI and 10,000 different variations—personalisation. But I don’t think that has too much to do with people being conscious of privacy and being afraid of that. I think that is viewed in a very different area.

I’ve presented about this and talked about this many times before. I still feel we are recording data for the sake of recording data. The majority of marketers would not understand or have a plan of why are we recording this. How are we using this to add value to the consumer?

And while that’s not being done in a nefarious fashion I do think the unplanned approach shines through. Consumers are not stupid. They understand this and they go, ‘you’re recording all this information about me; what are you actually doing with this? How are you actually building a better product, providing a better service?’

I think that’s where a lot of this frustration is coming from because they don’t see that and I don’t see that in my interactions with most brands.

Darren:

O.K. but in personalising the communications, if you do it clumsily it freaks people out. We’ve all heard the story about a retailer in the US (maybe it was Target) who suddenly began sending messages based on browser behaviour to suggest that someone was pregnant before this person had informed anyone else in the household about that pregnancy.

Christian:

That story doesn’t die.

Darren:

It’s become a meme. It lives a life of its own. Except that even on a personal level I’ve had people say to me, they’ve got Alexa (the Amazon voice search), they’re having conversations about a holiday and then they’ll suddenly see offers pop up about holidays when they haven’t actually said anything to Alexa.

Then we find out that most of the companies, like Amazon, Facebook, and Google have all being transcribing conversations manually, using people, just to check that the voice recognition is working.

So, it’s all of this lack of transparency that suddenly pops up in interactions with all of this technology that freaks people out.

Christian:

I think the main topic here is using data to customise and personalise an ad. And quite frankly, a consumer doesn’t perceive an ad as any value added. If it’s more or less targeted it doesn’t matter it’s still an interruption of what I actually want to do which is watch something, read something.

So, using data to customise and personalise an ad is not a value add in my mind or the consumer’s mind. And so, we’ve just got to change our mindset as marketers of why we’re using this data—to personalise an ad; not good enough. We’ve got to have another reason.

Can I use it to provide a better product? Let’s remind ourselves what marketing means. Marketing is price, product, place, promotion; a whole bunch more than just promotion, right? So, can we use data to make the product, price, and placement better? And that’s what I don’t see.

People are not using data to provide a better service or product; it’s always about promotion and we’re getting tired of that.

Darren:

It’s an advertising world. The business model seems to be completely focused on advertising but not actually customer experience.

Christian:

Right, it’s a cop-out though. I can’t be bothered focusing my energy on building a better experience; I’m just going to do the easy way out and throw more money at the advertising bit.

Darren:

Because that’s where they see a return on investment. Even though that return on investment is infinitesimally small because they pump more and more money into basically training the consumer, their customers, into how to ignore their very best efforts to get them to transact.

Christian:

I think there are great examples now for smaller, more agile companies who are doing that in a better way and they’re just focusing on building a great product or service. And people start talking about it. You don’t actually need any advertising. People will automatically start talking about the product if it’s good and that’s where all the effort goes. I like that.

Darren:

I think we’ve seen GDPR in the EU arise because that’s about the only way politicians think they can control this. They’ve had to legislate around how data is collected, what you do with it; you have to get permission (informed consent) to be able to do this stuff.

It seems to me we missed a big opportunity because there was a quid pro quo here, which should have been I will give you access to my data if you promise to make my life easier, more worthwhile and enrich it in some way.

Christian:

I think you’re right. There was a period where the industry had a chance to self-regulate in that regard but I don’t think we should be too harsh here on the industry and say we failed because in the end the reason why the GDPR is there and all this crackdown is now happening is because we’ve let some companies get too big.

There are 3 or 4 companies that have all the data and they have amassed a market-controlling position and I think that’s really what politicians are fighting back. It’s not so much the smaller companies. Funnily enough though, the legislation has actually made the bigger companies even stronger.

They’re the only guys who have the scale and the money and the budget to deal with these issues and properly support GDPR and build all the features that are required by the legislation and so it’s kind of backfired a little bit.

Darren:

Because the small companies struggle with being compliant. So, even being able to use the data they’ve collected to improve the customer experience becomes way beyond their capability because they’re so worried about being in breach of the legislation.

Christian:

Yeah, it sucks out a lot of their energy. And that’s why I think we’re seeing the political conversation changing right now to anti-trust again. It’s moved on magically from the privacy. Now we’re talking about anti-trust and breaking up because that’s what they’re really after.

They tried to work it one way and it didn’t work out and it’s moved on. That’s my personal theory.

Darren:

You mentioned earlier that your focus now is about collecting data within organisations and using that to allow people to be more efficient and effective. What could people be doing in their organisations if they had a way of collecting data about how they work?

Christian:

Maybe it was just me, running Datalicious in a very inefficient way with no idea what was going on in my business. But my gut feeling tells me that for managers and management in general, it doesn’t take much for you to lose touch with what your team is doing.

I had a few great examples of people only 2 or 3 steps removed from me and I had no idea whether they were doing a great job. It just never really bubbled up to me or I had to run to the next sales meeting or whatever (there was a lot of other important stuff going on).

So, I really feel we’ve turned a corner and reached a point where data is so readily available in a team context; I can see all the meetings you attend, emails you send, files you edit, Slack messages you post, all the DUO tasks you complete. We can see all that. It’s all readily available through APIs.

So, why not use that information to create some additional transparency in the organisation? What’s doing well and what isn’t doing well? Because, ultimately, I feel when people take a new job they are quite pumped about that new job and they want to do a good job. They want to contribute to the company; they want to change things.

But over time people get desensitised because we have this hierarchical management structure in place where only the guy at the top really knows what’s going on and withholds a little bit more information from everybody else.

So, it becomes really hard to contribute because you’re not really understanding the goal, the objectives and we’re trying to solve that through KPIs. But really what we want to do and what the employers are actually asking for is they want to see how the company is doing.

We tried that as an example with Datalicious and we showed the employees our PNL for a couple of months. Unfortunately, that was a bit of a mistake because they see the profit and that was the only thing they see. They don’t understand the risk associated with generating that profit. And so the only thing they do is ask for a raise, right?

Now, we’re trying to create a scenario where everybody can see the performance of the business: where are we going, what clients are performing well, which ones don’t. And we make that very visible.

It turns into a bit of a tribal culture where everybody knows what the common objective is and can contribute to that. That’s at least the idea we’re working on.

Darren:

It sounds exciting. Datalicious had many employees across how many markets?

Christian:

In the end it was close to 50 people across 7 markets.

Darren:

If you scale that from 50 people to 500 people to 5,000 people.

Christian:

You multiply the problem; it gets even worse.

Darren:

But this could actually help solve that because what we’re seeing is a lot of big organisations are doing things like trying to embrace agile; agile philosophies and ways of working because they’re realising that the traditional hierarchical approach (as you said a minute ago), that the guy at the top has the vision and everyone’s in their hierarchy and they get the information that they’re required to have to do their job.

If you democratise the performance metrics of the business, if people could see where value is being created or lost, where performance is multiplying the returns for everyone then you’re informing people to be more engaged, to be more agile because you’re using information to collapse the hierarchy.

Christian:

That’s exactly the idea. We’re trying to create a very simple metric, which we’re calling revenue per event. We’re tying all the different events together and we’re borrowing from our experience in attribution where it’s about distributing revenue across events. So, that’s what we’re doing in HR.

We’re taking all the events that happen in a team and we’re tracking them all the way through to the invoices that are being sent out for these events. And we calculate revenue per event. We can aggregate that at a team, client, geography level and graph that over time.

People may say that’s a little oversimplified but in the end a business is about making revenue. Forget about all the other KPIs and metrics. They just sprung up because we had no better way of looking at this, right? But if I could give you an idea at every layer of the organisation what the revenue per event looks like, that becomes a very powerful single KPI that everybody understands and can work towards.

Let them make their own decision. Let them come to work in the morning and look at their list of tasks they have to achieve and let them prioritise themselves. I’m working on this client over here–$5 revenue per event, oh there’s a task coming up $20 per event. I should probably be working on the $20 revenue per event task rather than the other one.

Let them. But it’s about giving them the right information and I think people would really welcome that and it could turn a workplace into what we like to call a tribe. Everybody has a common goal. They understand it, it’s simple and let them self-organise. We’re a little bit further away from making that a reality.

Darren:

But isn’t that what organisations that have embraced agile ways of working are trying to do. They’re trying to break down the hierarchy into teams of people focused on delivering particular data in a more agile way.

I’m saying the part that’s missing from that is that there’s no feedback in the feedback loop.

Christian:

That’s a very good way of putting it. Agile, the way it’s implemented these days, without that overarching information about what works well and what doesn’t is really just a way of making mistakes faster and correcting them, faster.

And there’s nothing wrong with that; I think it’s great. We’re using an agile way of building this product and it helps us; quick feedback, not working, change it, do this. But overall, in that agile environment, if you don’t have some intelligence in there that helps you prioritise and judge whether what you’re doing is good or not then you’re still quite ineffective.

Another key point is it doesn’t contribute to the overall morale and motivation of the business. That’s one of the key things we’re trying to do. We’re not trying to turn this into a big brother that you use to audit whether someone is at work long enough.

Darren:

That’s where my mind went when you first mentioned collecting data within organisations because we’ve all seen those 1950s, 60s films where the time and motion expert would turn up with the clipboard and stopwatch and work out how long it took you to do task A and if you did it this way it would be half that time.

But you’re not actually talking about the data of efficiency; you’re actually talking about the data of organisational performance.

Christian:

Yeah. There is obviously the potential that a tool like this will be abused like that. We’re spending a lot of time thinking about how we can avoid that. One of the ways is not turning it into a bean counter tool, not reporting individual down to the penny or the minute; it’s about graphing major trends, something going up or down from week to week or month to month.

There are ways to navigate that but in the end you need a business that has a certain attitude. It wants to be more open with its employees, create a more streamlined hierarchy and then I think you can make that work.

We’re living in a society where privacy doesn’t really exist anymore. We’re fighting back with legislation, in lots of different ways but ultimately, if we’re honest with ourselves, privacy has come and gone.

Darren:

I think people are trying to cling to the idea of privacy because they have not felt the benefits of giving up on it. They have only felt the downsides of giving up on privacy. When people feel out of control they want to get back in control. How do I get control where I feel that my private information, personal information about me is now beyond my control because it’s now belonging to someone else who can use it however they like?

So, in some ways, trying to legislate privacy is incredibly difficult. It’s a ham-fisted way to answer the problem, which is privacy doesn’t exist. Every time I interact with almost anything data is being collected about me.

I remember in London they said there are more CCTV cameras in London than in any other city in the world. That meant that as you were walking down the street data was being collected about where you went anyway.

Your mobile phone is a tracking device whether you like it or not. And it’s also a microphone that can listen in on conversations. We see those spy films where they smash the phone and remove the battery. These are all, in some ways paranoid but in another way realistic views of the world. We live in a world where we are being observed.

So, the question should be not how do I stop that happening because the only way to do it is to go off the grid, go and live somewhere where there is no internet, no satellite surveillance and live that primitive life. Or start to demand a return on that investment.

The investment is I’m giving up this data about me (which you’re going to take anyway)—all I’m demanding is a return on that investment.

Christian:

I think we’re starting to see that in the US. There are some start-ups in the US that are now allowing consumers to monetise their data—contribute what they know about themselves and then sell it off to individual brands who are happy to pay for it.

Funnily enough, the price being paid for that is actually not very high. It’s not being valued at a very high price. We’re going to have to overcome that issue because otherwise the consumer is just going to feel cheated again.

Darren:

I think that as soon as you turn it into paying me a dollar amount. I’m a Qantas frequent flyer so I’m happy to give up that information because I get, depending on my loyalty levels, a certain level of recognition. Being paid an emotional return is much more powerful.

If they turned around and said we’re going to give you $2 off your flight or a free drink—well that’s ridiculous because you’re turning it into a monetary value.

Christian:

It’s always going to fall short.

Darren:

Giving people recognition, that’s ultimately why brands are wanting to collect this data about their customers because they want to give them an experience that makes them keep coming back and spending more and more money with you more often to make the business more profitable and not go to the competitors.

So, no, don’t pay me for my data but give me a level of experience that’s commensurate with what I’m giving you.

Christian:

I think you’re spot on and I think that lesson has been learned very quickly by these companies. They don’t tend to stick around for very long. They pop up every one or two years, give it a go then just quietly fade into the background because you can never pay the consumer enough to make it feel like a gratifying exchange.

Darren:

But what I like about your focus on using data and analytics to make employees feel more part of the success of the organisation (I’m purely coming at this as a marketer) is that I can see it will overcome one of the big problems we face in business, which is all these areas of specialty all trying to play land grabs.

The sales team are arguing with the marketing team, operations with someone else, finance people argue with everyone, no one wants to talk to procurement but if you actually got this to work (and I’m drawing on the fact that when we first met one of the things was media attribution and attribution models based on data) is that you could start to build into this, once you’d got enough data into the system for the company, attribution of value.

Because it wouldn’t be about $5 for that event or $20 for that event; you would start to see the organisation as an organism where all these parts are making a contribution to the outcome.

Christian:

Absolutely. I think it’s time we have a computer or algorithm decide that for us. We can do that; the data and technology is there and that’s what we’re trying to do. Take all the data out of it and give a report saying this team has contributed X revenue, this team has contributed Y revenue.

We’re going to do it in a way that protects people’s individual privacy but it will give you information to make a better decision. We’re borrowing heavily from our past experience of media attribution because it’s pretty much the same thing. Tying revenue to ad impressions versus tying revenue to individual team actions is a very similar concept.

I did consider becoming a barista for a while but figured it was better to stick to things I know and have experience with.

Darren:

I’m sure you’ve drunk enough coffee to think you’d probably be a good barista.

Christian:

It would be disastrous.

Darren:

We’re still seeing those discussions happening between sales and marketing where sales are taking the value of the bottom of the funnel: last-click attribution, in the real world, and marketing, because there is no agreed attribution for their contribution they’re moving down the funnel to play at that same level.

Christian:

The point where it always went wrong with attribution projects and the point where it’s probably going to go wrong with this new HR analytics tool that we’re working on is if it’s being used for bean counting.

If you’re trying to figure out how many pennies for this person or campaign—who cares? You don’t run a business like that. You’re trying to use information to find big trends. Am I growing in this space? Have I generated a big spike with this campaign or activity? Is this client generating more revenue versus the other one? Is this team more efficient than the other one?

Anything that shows me a significant trend, whether it’s positive and I can exploit it and try and create the same trend elsewhere or it’s negative and something that I have to fix—that’s what I’m looking for. And marketing people make that mistake as well. They turn it into a bean counting tool rather than looking for changes and using it as a tool to implement an effective test/learn strategy.

If you don’t do that it goes wrong. And we still see that. It’s a lesson everybody has to learn at least 5 times before they really incorporate that.

Darren:

It’s true. I think so often we’re working in a cause and effect universe. We think that if you do something and you get a result, that every time you do it you’ll get the same result.

Christian:

Yeah, correlation, causation—not always related is it?

Darren:

No. In fact, if you look at complexity theory, and most human systems that we create or exist in are complex; they’re not simple, there’s very little cause and effect. They’re actually incredibly complex systems where at best you can notice trends but you can’t actually get to cause and effect.

Economics will point to particular things but they always do it after the event. They’ll always try and make sense of the changes that happen in economics. Traffic flow is another one that’s incredibly difficult because it’s complex.

Christian:

I think that’s one of the key aspects: providing information in a timely fashion. Old-school mixed media models done once a year happen after the fact. By the time you get the results and use that to plan your campaign for next year, things have already changed. That’s what we tried to change with media attribution. We tried to create a real-time reporting mechanism that shows you every day how things are going.

You have very up-to-date information on what’s working and what’s not to then implement an ongoing test and learn culture in the organisation. Just to facilitate this test and learn thing, not to encourage a bean counting exercise.

That’s what we are trying to establish with our HR tool; it’s not meant to bean count where people spend their time but unearth big macro trends in the organisation and doing that in real time rather than at the end of every week, quarter or end of year when you look at your PNLs, so you still have time to make a decision.

The big overarching strategy, and I’m not sure if we’re ever going to get there. Initially, I think it will just be a dumb reporting tool with some automated alerts but our grand vision is to build an AI engine that assists people in managing their business.

Building an AI is a very complicated process and you really need a common goal to optimise to and that’s not present in most organisations because the environment is too complicated. You’re trying so many things to make your business run. If you give an AI to learn all these things it’s probably not going to happen.

We’re trying to simplify that and come up with this new metric; it’s called revenue per band, and that’s the single goal that we optimise the business towards. Maybe down the track, profit per band, if we can crack that but the point is it’s one metric.

We’re going to try and accumulate that across several clients (probably need a few 100 for that). Then we have a true crack at building an AI designed to assist businesses in managing the day to day business. Wouldn’t it be amazing if in your business there was a little computer sitting there telling you the things that go wrong, giving you advice on how you should be managing your business?

You can focus on your core competency rather than having to deal with all these other distractions which suck up most of your time in the end.

Darren:

Absolutely but also tells you where to focus your efforts and energies.

Christian:

Gives you guidance in priority.

Darren:

So you can prioritise your efforts. But I like the idea of the democratisation of that within the organisation. I have heard people talk about how AI is for the CEO or CMO or CFO but the idea of actually providing that real-time feedback loop that is allowing the organisation to align and perform in the areas that are going to give the biggest return or benefit.

Christian:

Yeah, The solution to getting around the big brother and privacy thing is to watch the granularity – how detailed you provide the information and number 2, make it absolutely transparent, everybody has access, it’s not a management tool. Everybody down to the smallest employee can have a look at what’s going on and I think that creates buy-in.

It immediately eliminates the idea that we’re trying to do something nefarious; you can see what we’re doing.

Darren:

Do you know many organisations that have truly embraced agile ways of working that are not in the software development area?

Christian:

I’ve never really asked myself that question. I’m really going out on a limb a little bit with this new venture. I’d just assumed that other people are just as inefficient in running small to medium-sized businesses as I was with Datalicious. Hopefully I’m not the only guy who had these issues.

I’ve had a lot of meetings over the last few weeks coming back to Australia and this is exactly what I’m trying to find. Companies that have that problem will find this approach interesting and running a private beta and playing with a tool like that. That’s one of the key things I’m trying to achieve at the moment.

Darren:

You hear of so many organisations going through a transformation to become more agile but talking to you today, the thing that’s missing is that real-time feedback loop because you can work in an agile process but unless you’re actually seeing benefit, a sense of progress; I think that’s why agile works so well for software development because it’s about test and learn, test and learn.

Imagine, when you reduce that to a standard business process.

Christian:

Imagine 3 different agile teams. In the current setup, whether these agile teams are working in the right direction or not is entirely dependent on the manager of the agile team.

Within that agile team they will figure out sooner or later or quicker the best way to conform to what that manger wants but what if that manger is wrong and working on the wrong thing? What if he’s not communicating insights well enough that everybody knows and understands what the true problem is and can effectively contribute insights?

What if all of these 3 agile teams are not dependent on their manager to find out what the business priority is? Everybody in these teams understands exactly or a lot better what it is that drives the business, that needs to be improved and everybody can come up with ideas and help prioritise their own agile process.

That would be a lot more efficient. So, I think that’s where the breakdown is in the current agile approach. It’s a good step forward but it’s not giving all the participants in the agile process all the information they need to do a really good job.

Darren:

Christian, it’s terrific to catch up with you. We’ve run out of time already but just before we go; how long before this brave new world will exist?

Ideal for marketers, advertisers, media and commercial communications professionals, Managing Marketing is a podcast hosted by Darren Woolley. Find all the episodes here