2022 – the year of privacy and FOMO / Data for bluffers #3

20 December 2021

What’s going to happen to privacy in 2022? Will data FOMO dominate? And why will small data rule the data conversation? Listen to our predictions of how data will impact growth teams in 2022.

All this and more in episode three, presented by Tom Ridges and Dr Ed Barter.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

Tom

Hello, welcome to another edition of the data for Bluffer’s podcast. I’m Tom, the CEO of GD labs. And as ever, I will be joined with ed ed say hi, hi Tom. This week as we are ending another year, me and ed thought we would look forward to the predictions for 2022 for data. Uh, and we talk about quite a range of subjects. We talk around privacy. We talk about phrase, which I, I now love, uh, data FOMO, and we auto talk about how small data is going to drive bigger impact to, um, to more businesses in 2022. I’ll be quiet. I’ll hand you over to the conversation. Hope you enjoy it, and I’ll, uh, speak to you later. So, ed, I noticed you’ve got your Christmas decorations started to spring up in the background, which means we are at the end of another year, which I think is a good time for us to chat around what are gonna be the trends in data for, for next year.

Ed

Luckily we, uh, we didn’t make any predictions last year, so it’s not the time to look back and see why we went wrong

Tom

Or, or we can just say we, we were a hundred percent right. That,

Ed

That recording didn’t get out. Unfortunately. Yeah, there’s a, there’s a few things that I think will be in important in 2022 and trends that are happening already, but will continue and will become much more prominent. The first of those is something that we, we talk about a lot and that’s sort of privacy and also awareness of the extent that technology is kind of gathering data in our lives. And also the extent that people are using that data to, or in our case, you know, target adverts, but services that people work aren’t necessarily aware about.

Tom

Do, can you break it down? Like it’s, you know, privacy in data is quite a big subject. So there are some kind of key things we should be looking out for next year or thinking about, you know, getting our heads around, how we’re gonna deal with as businesses. How would you cut that down? The

Ed

Way that privacy is always developed in data is that whilst there are sort of legal and reg litera pressures, ultimately that is coming from the public, you know, we kind of got distracted by the pandemic and now people’s attention is actually moving back to sort of questions of data and privacy and sort of the invasion, I would say to a certain extent of technology into our lives and how invasive they are. Yeah. Recently there was a case in the UK where someone managed to Sue their neighbors over the storage of video footage from their ring doorbells. And this isn’t necessarily like a direct privacy concern, but it shows that people are coming more aware of how these things are being used and how much data is being gathered from our

Tom

Lives. It’s also the way it makes into main through media as well. I’m, I’m, I’m watching a show on TV called succession at the moment, but what was interesting is he made a qu about having Alexis in his house. Enough, people are already listening. I don’t wanna add Bezos to that list. So it’s really interesting that this, you know, this, this, this narrative about tech and privacy is now is now just standard commentary and TV shows.

Ed

That’s a really good example where the awareness is now driving people to start asking, do they want to do this? I think from a legal perspective, what that norm, where that normally ends up is we start talking about consent. And a lot of are you uses of data break down. If people let’s be honest, get given genuine consent mechanisms because people will opt out an example that we’ve we’ve, you know, has been spoken about a lot over the last year is the removal of third party cookies from Crow. And there are some Google is developing some replacements how good those replacements are, is to be seen. But one particular worry about those replacements is you’re gonna have a centralized consent mechanism, most likely, which means that someone will toggle cookies on or off or D replacement for cookies, sorry, on or off once. So you don’t have that chance to circumvent it every time they hit your website. If we look at

Tom

What’s happened with apple, right? And they’ve got location tracking in the late versions of, of iOS, most people, if you look at the data they are hitting to opt out, when it’s easy to opt out, as apple have made it, people opt out it’s, it’s the consent mechanisms at the moment that are just so complicated to opt out that you just hit hit allow. So to your point, I guess when there’s a centralized mechanism and it’s fairly easy to vigor opt out by default, then yeah. Think things become tricky for, for cookies.

Ed

If people become aware that every time they’re having data collected, they’re gonna have the option to opt out. People might start looking for it as well. So I think this could be a

Tom

Broader issue. Yeah. And the other one, I think’s really interesting that, that doesn’t get a lot of air time is around IP addresses being classed as hi. When we listen back and realize we’ve used an acronym, I’ll jump in and explain what it is. So in this case, PII is any data that could potentially be used to identify a specific individual IP addresses being classed as PII. You, you talk about these other mechanisms that people might fall back to and, you know, IP addresses get get about, but actually that’s, you know, from a compliance point of view, that’s not there. And, you know, we we’ve seen cookies going away, but there’ll be mechanisms to, to block that other data soon as well.

Ed

We kind of seen it already this year to a certain extent, but next year it will be the crystallization of what is the post cookie landscape on Chrome. So Google’s got a few components, it’s privacy sandbox as they call it where they’re developing sort of alternatives that sort of protect privacy. They had some trials this year that haven’t necessarily gone brilliantly. It’s not like they’re gonna replace cookies perfectly. I believe there’s still ongoing legal trouble for that because the effect of this is your, the ad targeting will basically be done by Google. And then it’ll sell the targeted cohorts out the back to the advertisers. And obviously this causes big competition concerns because it’s basically saying that Google is the only company in the world that can do ad targeting what I think we’ll see to a certain extent. And what I really hope we see is the transition from, okay, how do we basically replicate cookies? So how do we do as close as we can to cookies, but within compliance and actually say, what will we doing with cookies? And what is the compliance and privacy sensitive way of doing that thing? If you were using cookies to find new customers, okay, how do we find new customers without using cookies, as opposed to, how do we get a system which kind of looks good and is as close to a cookie as possible, and then use that to find new customers?

Tom

How, how do we actually look at this and do it properly for the long term, right? Because I think people are waking up to this and, and, and it’s, it’s only becoming more aligned to people’s requirement around privs. So just trying to, trying to make tactical changes to get around it in the short term means you’re gonna be making the same tactical changes in 2023 and 2024. So it’s, it’s a good time for the industry as a whole to look back and say, well, how do we make, how do we make long term range that actually solves what we’re trying to do, as you say, you know, if it’s trying to find customers or whatever you might be trying to do, but also aligns long term to the customer and the consumer requirements,

Ed

The solution to the problem of finding new customers using cookies is not necessarily the same as the replacement for a system that you are using. For example, to use, you might be using cookies to prevent someone seeing ads more than three or four times a day or the same ad. So the system that you might put in place to prevent that happening again, almost certainly will be a different system to the system that you’re using to find new customers, even though they both used to be done by cookies.

Tom

Yeah. Interesting. There’ll be a lot of innovation, I guess, in that space, in the next, next day, 18 months coming out,

Ed

We certainly hope so. So people will be, should be looking for solutions now rather than looking for solutions, you know, at the back end of next year.

Tom

Okay. So privacy is our, is our number one prediction, where do we go next? I

Ed

Still believe that, you know, data and data science is, is growing it’s booming and it will continue to, and I, I see there being more, more FOMO in data science. So at the moment, it seems like everyone is scared of missing out on the gains that can be made by using data. Now we’re already seeing, you know, there’s a lot of stories coming out about the lack of success of data science projects, particularly in big businesses. This is often driven by the fact that goals aren’t or, or objectives for data science projects aren’t particularly clear. Yep. And people are really driven by the desire to do something rather than the desire to do a particular thing. And I see that continuing in into the new year, unfortunately.

Tom

Yeah. And, and, and I think what gets interesting is I’ve seen on the top end of the scale, lots of large organiz that can pay big salaries, hiring lots of data science people, and these guys and girls not having enough to do, you know, so I’ve spoken to loads who are almost twidling their thumbs, but the problem is because it’s a lot of these big organizations that are paying them and there’s not enough data science resource it’s driving up the salary. You know, there’s, there’s eye watering sums in, in the states. Like you hear stories of, of Google hiring grads for like a million quid or a million dollars, right? The top end, we’ve got these a lot of big org inflating the price of this resource, but not using them fully. And then at the other end, you’ve got smaller businesses who don’t have any data skills in the business, but also can’t get close to affording the salaries that these big tech businesses can deliver. So, you know, I I’m, I’m always, I’m always intrigued from the sideline, how that, how that plays out. Cause there’s this big desire to do more with it. No one wants to miss out. You’ve also got some big companies who are, who are idling with data science resource driving up the cost and the majority of businesses. So they can’t, they can’t play the game. For

Ed

Example, you know, I think it was Shopify this year. You know, they pledged to employ 2021 data scientists right. In 2021, which I mean is marketing driven recruitment. if you’ve ever seen it, but ultimately, you know, how much value can you possibly get out of that? And they might be able to try lots of things and they might find the best model for doing what you want eat to do, but you know, what, what cost of getting there. And especially if that’s going to be an ongoing commitment to have that many data scientists, right. You know, you’ve got people like Google doing, you know, amazing things, cutting really cutting edge. But the two, even a matter is, is that the vast majority of businesses are right at the other end of that data revolution. And there’s sort of small steps they can make that would actually give much, much greater gain.

Tom

Okay. So we we’ve had, we’ve got, we’ve had privacy, we’ve got fear of missing out. We’ve got FOMO and some of the issues that could cause positive and negative impacts within that, let’s round off with a third. What’s our, what’s our top three predictions for data in 2022.

Ed

So I, I dunno how much of a prediction this is, but this is definitely what I fi think is probably the most exciting area of research really. And that’s looking at like small data and what’s sometimes called tiny ML, which can be a very specific thing. But here I’m saying like generally the idea of, of doing machine learning, but with small amounts of data and smaller machines, you know, not big cloud computing problems is

Tom

Small defined, you know, do you, do you cross a threshold and you, you go from a small data problem to a big data problem.

Ed

No, I don’t think, I don’t think it is defined in that way. So most of the re search in this space, in the small data, tiny ML space is driven by or driven into that space by some sort of restriction. Right. So the restriction could be, for example, looping back to what we were talking about at the start could be privacy. Right. Okay. A very, very common example that of people have encountered is the, you know, Alexa or okay, Google wake up words, your, you know, your, your Alexa is listening all the time. Yeah. And then to prevent the need to send that information back to Amazon, you can run machine learning in the device itself that does the detection and decides whether it needs to wake up. And then when it wakes up, it starts to stream, okay. Back over the internet and do the more complicated machine learning stuff.

Ed

So, and that, that, and that’s been driven by kind of two things. One is, is a need for privacy. The fact that if you are, if you were just passing all that, all that data back to Amazon, people would probably be a bit more concerned about it, but also it’s, uh, it’s also to do with speed. So the fact that you want, you want it to be able to respond without having to send all the data down through the internet, to a server down in Amazon, which calculates yes, I need to turn on and then get a message back saying, turn on an example, that’s becoming a lot more prevalent is in self-driving cars where they need to make decisions quickly. You know, you don’t have time as you’re driving at a wall. It’s a hope that your 3d signal is good enough to send that data back to yeah. A central, a central, uh, solution where it can run a machine learning model and five minutes later, come back a yeah. That was a wool yeah. That you crashed into there. Right. So that, that I, the, the kind of need for speed in, in processing and the need for privacy are kind of the two key drivers or main drivers in this, but from my point of view and why I think it’s exciting is that the techniques that are developed for those problems are centered on getting as much information and as possible out of a small amount of data with a relatively lightweight model,

Tom

Which I guess links back to what we’ve talked about previously with, with FOMO in the, a lot, a lot of the problems before have been solved with, you know, air quotes, big data. And I think talked about this previously, right? That there’s often a lot of anxiety. When you talk about, you know, I call it data anxiety. When you talk about a data project, people start trying to think of, well, this data could be in their heads, how they’re gonna get access to it. Is it hidden behind the sofa? Whereas it gets, if you’re working, if you’re able to extract maximum amount of value out of small amounts of data, the opportu for more people to get on the ladder, if you like or starts to address some of that, that, that FOMO piece, right. It makes it more accessible to people to get moving quicker, prove results, prove value to the business, and then tackle, you know, bigger, more complex problems once they’ve, once they’re more established or, or I think again, as we discuss either week, whether you actually need to a need to tackle that bigger problem, you know, sometimes I think it’s a, sometimes people go after big data looking for a problem that they don’t necessarily have in, in some scenarios

Ed

Exactly like the, the, the, the research that’s driving in this area really helps smaller businesses. For example, get into the data space. It else people dip your toe in and for a lot lower investment, start seeing the benefits and understanding how your customers interact with that. And whether that is something that is maintained whilst your customers are still interacting with it. And it also makes the whole process a lot cheaper at the end of the day as well. Right. You’re not relying on big cloud computing mm-hmm to, to deliver these solutions also from, from our point of view, as a business that, that sells sort of a, a data science service, it’s a way of us helping more people more quickly. Yeah. So we are able to build, you know, smaller models for people who are sort of just getting interested in data. And also we are able to deliver insights almost instantly, as opposed to waiting for them to go and find their data. Yeah. You know, gathering all of it up and then find or expecting to find the best, the most value we possibly can to justify the cost. Yeah.

Tom

Hun hundred percent. And actually when I, when I speak to customers, you, we speak to a marketing or a growth team. Uh, you know, and we say, look, we can, we can help you find more customers quicker, but we need your first party data. You, you can almost see the anxiety building in their head of like, well, how am I gonna get hold of that until you explain that? No, you know, we, we work on small data and we only need these things and you, you can almost feel them breathe out and go, okay, that that’s doable. You know? Cause I think everyone’s been in this world where someone at some point has asked them to get some data and it’s been an absolute nightmare to find all the different pieces you need.

Ed

I think that’s, yeah, that’s a problem in all businesses. And especially now, when the ICO makes a judgment, it seems to me,

Tom

The ICO is the information commissioner’s office, which in the UK upholds information rights in public

Ed

Interest, it seems to me that a driving force behind that judgment and the punishment is to do it’s to make sure that people don’t just chance their arm and hope they either don’t get discovered or, you know, if they do, then the privacy case is worth the value they’ve got out of that. But that does make people hesitant to do anything that might challenge privacy. And so being able to work with small amounts of data, it means that you, you don’t to pay a bank of lawyers for a week to check that, that you are compliant all on all the data fields that you’re planning to use. You might still have to get a lawyer, but it might only take them an hour to look at the five fields or whatever that you, that you, that you want to use.

Tom

Interesting. So I think we’ve got three top predictions, uh, privacy and awareness and how that’s gonna in the next year. FOMO. I love that data FOMO and more focus on tiny data and ML. So from my side, I think we’ll wrap it up. I’m probably gonna go and have a mince pie as it is December and officially, you can eat mince pies in December and as you can only eat them in December eight, a lot of them, uh, any parting w them or forecast crystal ball bombs from you before we go, ed, I would, I really

Ed

Wanna say I have a crystal ball. I think these are, these are things to, to look out for over the next year, but I’m sure that as data as data starts grows, um, we’ll probably be more

Tom

Surprised than we think. Cool, good to talk you all right. Speak soon. Speak to you soon. I hope you enjoyed the conversation me and ed had, and you know, we’ll come back in, in the end of 2022 and, and review our predictions. We’d love to hear from you. What, what do you think, do you think that what we’ve talked about makes sense, do you think that there’s gonna be bigger trends that are impact data, particularly in the, you know, the sales, the marketing and growth, and we just talk in rubbish, um, you know, let, let us know. Um, but in the meantime, if you enjoyed the podcast, like subscribe, share it with people who you think will benefit from learning about these concepts in the sales, marketing, and growth space. And we will see you again in two weeks time.

Friends in conversation | Herdify

Sign up to the Herdify newsletter