Good afternoon.

Welcome to this webinar from 51Degrees, where we're going to be looking at publisher and advertiser and ad tech uses for the 51Degrees technology.

We're going to start with a little bit of background for those of you who aren't familiar with all the real-time data services we provide, then we're going to look at identifiers, which is a new feature that we're adding to the product set.

Then we're going to look at AI theft protection, how these can be used together to protect copyrights protected content.

Then we're going to be looking at revenue optimisation.

And finally, a little section on migrating to 51Degrees.

So we're packing a lot in.

Alex is then going to join me, and we can take your questions.

Let's get to it then.

So 51Degrees.

Think of us as taking raw materials, that's user agents, or IP addresses, things like that.

And then refining that data.

That's what happens in our data centres.

And then providing you the refined ingredients.

We don't provide finished goods.

Finished goods is what you do as our user as our customer.

And we're very much going to be focussing on those refining ingredients, what we can provide to you before we then talk about how you can use them.

So, let's start with a little bit about the company.

So we train our device detection solutions on 1,200,000 device combinations.

That's halved in the last couple of years, as our algorithms have improved.

We're ISO 9001 and 27001 certified, we use real people, real devices and vendors, to research information, no AI is used in the data research and integrity part of the business.

We have over 94,000 profiles of devices, and that in deployment works at over 99.9% accuracy in the real world.

With data updated daily.

We publish information about the quality of our data, you can see that online.

You can see how it's changing over time.

And also some information about the devices that we add a lot more TVs more recently in media hubs than say 4 or 5 years ago.

So let's have a look at some of this in action.

So on our website, we have testers for all the different raw ingredients that we provide.

So we've got the IP tester page here.

This is taking my IP address, we can see that the service area is quite large, but I'm most likely to be around the London area, which is correct.

We can see a little bit of information about the network type, the connection type, human probability, some information about the country, so yes, I'm in the United Kingdom.

You get the WKT for the area, and location confidence coming back as well, in this case, high.

You would use high, for example, if you wanted to make a consequential decision, but maybe if we were less certain, medium or low, you might not use the data. Information on time zones, and then some general information like device diversity.

This is quite a widely used IP address, part of Carrier-grade NAT.

So we see quite a large diversity.

So now let's try a user agent.

This is another one of our refined ingredients.

You can see some information about the device I'm on, but if I want to try a different device, like a Samsung, we can see information about that.

We can look at the evidence that was used.

This is fuzzy matching, so we can see some of the strings and characters that were important. Information about the screen, the date that the different components were released onto the market, whether they're the latest ones or not, and we can see some more general information as well, about the supported bearers, just loads and loads of information.

You can play with these.

That's why we put the testers up online.

If we look at type allocation codes, These are the first 8 digits of a mobile devices, imi number.

We've got some we've prepared here in the tester, so you can stick those in, and you can get information about the device.

Obviously, we don't return the operating system or the browser, but we can return some information about the physical device.

And then finally, if you want to see all of this together, we've got your device. Where we're actually coming back with a device that is currently being used, where the 3rd party cookies are enabled.

We're going to come back to that later.

That's quite important to the value that advertisers place. On an opportunity to advertise.

So those are our 4 testers.

Very easy for you to play with those raw ingredients that we provide device location, human network, et cetera.

So now let's move on.

We've got this new feature we're adding around identifiers.

And really, if we look at the market for identifiers at the moment, we have lots of contracts between the user and the advertiser that are contain lots of common features, but are not open.

They're not readily understandable.

So, the user enters into a contract with the publisher, the publisher with the SSP, the SSP with the exchange, the exchange of the DSP, there's obviously other parties involved as well, and then the advertiser as well.

Each of these are different contracts.

So what has to happen when we have an identifier at the moment, or what happens at the moment, is the ID vendor comes along and they have a contract, and they enter into that contract with the publisher, and they enter into that contract with the DSP, and any one in the middle, just sees encrypted data, so they can't really do anything with it.

So this is using encryption in order to protect that identifier, to meet data protection requirements.

Now, what 51Degrees is doing is using a single contract for the purposes of protecting the identifier.

This isn't replacing the contracts between the different parties for things like payment, and terms, and performance, and indemnities and things like that.

It's just covering the identifier.

So what 51Degrees can do is provide that identifier to the publisher, at the very beginning of the process, obviously having confirmed with the user that they're happy through a CMP or some other form of dialogue, and that contract, then follows the identifier through the supply chain.

So we don't need encryption.

The contract prevents abuse.

So we're using legal means, in order to protect the abuse, the ID vendor needs to sign the ID for integrity in that case.

So, in this case, it's us as 51Degrees, but of course, anyone can enter the market and use the same principles.

And publishers get all of this from the 51Degrees solution in one go.

What we need to do, though, before we can provide anyone, this 51Degrees identifier, this 51 did, as we call it, for short, is we need some extra checks before we provision the service to our customer.

So there's two things we need to go through; know your customer.

So you can't just rock up on the website and get a normal license key for our cloud service.

We have some enhanced, know your customer checks.

And then, of course, we need some indemnities backed by insurance, in case there's a breach.

So we're using contract law with some extra features at the beginning of the process with our customer in order to protect the data.

So, the benefits are, it's reducing the number and the complexity of contracts for common purposes.

All those specific contracts between parties, they can still exist, but for this particular piece of data, it becomes simpler because a common single contract is being used.

It increases transparency through the supply chain, because there's a common contract that's used to protect certain types of common data, and ultimately, it's reducing complexity, both technical and business, but particularly that technical complexity associated with the encryption.

So what we're doing through the data life cycle is we're taking information, like user agent, latitude, longitude, IP address, people's preferences, we put that through the machine learning, that refinement capability at 51Degrees, and then we provide those refined ingredients.

That can be data files for on premise consumption, can be access to our cloud service, which you saw earlier in the testers, or we can actually provide you a docker container that you can run in your environment as well.

So those finished goods, or analytics, optimisations, fraud, advertising, of course.

So we obsess over those refined ingredients, providing all of them to you, in one single contract, if you wish.

Okay, so we're going to come back to the use of those identifiers in advertising in a little while.

Let's talk about robots and AI theft protection now.

So, in the field of uh, say washing or the offline world, we used to come with these common contracts, the kind of contracts I talked about earlier with the 51 did.

So washing instructions for clothes, so they can last longer, you can look after them.

Garment vendors don't have their own iconography.

Every garment vendor uses the same set of icons effectively so that we can protect our clothes.

We also have these common contracts that are used in the real world, rental agreements, for cars or apartments, or holidays, et cetera.

All these common contracts.

So why don't we apply these to the online world, the digital world?

Well, data labels is the way to do that.

We had a label to copyright protected content in the case of content protection, or it could be indeed embedded into the asset as well.

So we call this concept, data labels.

What we've done for the purposes of protecting content is create a day to label with Movement For An Open Web.

So Movement For An Open Web is a not-for-profit that I started with Tim Cohen in 2020 to try and advocate to bring the web back to what it might have been once, a long time ago.

You can find out more on the Movement For An Open Web website.

We have sessions that do a deep dive into this technology that you can just contact us about and we can put you in touch for those sessions.

What MOW’s label does, is it's kind of like a permit for parking, a parking, sorry, parking charge.

So what happens is when a accessing party accesses the website, the terms and conditions for that website, using this common data label, um, that protects the content so that it can be used for search, it can be used non-commercially, um, for viewing, but it can't be used for other purposes, like ingestion into an AI.

So what happens is it creates a debt that is waived if the terms are complied with.

So kind of like parking, you don't pay necessary to park at a certain car park, maybe for retail.

But if you stay for too long, then you end up having to pay the debt.

It's exactly that principle.

Anything else is going to be a breach.

So, say used in LLM used commercially, and that can result, or that results in a £500 debt, that amount can be changed by the publisher or the website operator.

And the debt can be collected via the UK County Court, which is comparatively quick compared to copyright or high court cases.

So what the Movement For An Open Web team have done is combine features of law and contract to provide a method effectively to protect people's content, publisher's content.

So most created, one of these contracts, 1000s, 100s could be created.

I doubt it's going to be 10s of 1000s because out of practicality, globally, having a small number of common contracts makes sense.

But the market will decide there'll be competition between these contracts.

So they're easily deployed into the technical standards today.

So with robots.txt, data labels has this thing called a terms document locator.

It's like a URL or a web address, and it can be added into the robots.txt file.

We prefix it TDL, terms document locator, and then you place the URL for those terms.

So Movement For An Open Web are hosting the Search-Only Contract for Web, SOCW.

At the moment, and you can add that to your robots.txt file.

You also want to update the terms and conditions to provide information about this license.

You do that through incorporation.

It's kind of like I was saying earlier about the 51 did, has this contract that flows through with the actual data, that would also be incorporated via reference into existing commercial contracts.

Essentially, it's essential that it's the same document, so that the AIs can understand it.

They're not overwhelmed by lots of different contracts, and each AI is then free to decide how it wants to behave when it sees that TDL, that license.

So you can see some of this in action for robots.txt.

We work with usages rather than crueller names. Just too many crawlers out there for normal people to be able to understand all the different uses that they put data to.

So we classify them.

That's one of the things we do.

And that refined ingredient that we provide to you means you can easily take, make use of that.

So we provide this free service on the website, 51degrees.com/robots-txt, and you can use that and you'll get a weekly update of a robots.txt file.

For those of you who simply want to deploy it straight into your website, you can do that through our cloud service.

Again, if you want to find out more about that, contact us and look on the website to find out more details.

So, we have the Movement For An Open Web contract.

It's like a signpost, when you come to park your car, and we make it really easy for you to keep that up to date so that you can signal to anyone coming to your site, but particularly the robots, particularly the air eyes, what they're allowed to do. And what they're not allowed to do.

But what about enforcement?

Where is the parking attendant effectively?

Well, that's where we come in as well.

We can help you there.

So you have your application.

You can deploy 51Degrees into that application, so if it's WordPress, you can just use our WordPress plug-in, or you can use the cloud services or the open source software that we have behind any of the features that we're showing you today or that you can find on our website.

You can put that into your application.

And you might configure that based on those usage parameters.

So you can say, yep, I'm happy with search, but I definitely don't want AI training.

So when the web session starts and the request comes into your web application, that filter can be applied to allow or reject.

So you could send the AI to a page for AIs to say how they can get a license.

So what 51Degrees is doing is providing real time enforcement at the application level.

It actually sits within, can sit within your application.

Of course, AI training, invalid traffic, for example, certain conditions you're not happy with, you could put the same filter effectively into the application, so disable the advertising, if certain conditions are met, or not met.

So just some examples of how you might configure these ingredients for invalid traffic.

You could look at the browser version.

Is it the current or the sorry, is it not the current version of the browser?

Maybe it's older than 3 months.

That one might be a sign that something funny is going on.

Most browsers update pretty quickly every 6 weeks.

Human probability score, if that were lower than two, for example, almost certainly not a human.

High frequency of requests.

So again, using those identifiers I talked about earlier, you could use them to look at the frequency of requests and maybe introduce some sort of throttling there as well. Combination of the browse of the OS and device that has low probability.

You can use the ranking values that we provide to identify that, which is low confidence location information like I showed you in the tester earlier.

You might not be comfortable serving advertising in those situations.

But we provide the raw ingredients.

How you use them, the recipes that you take those high quality, refined ingredients, that's up to you.

Our job is to provide you the very best refined ingredients.

So recipes, how you deploy them into your secret source, for example, that's up to you, but we really, really want to work with you, to help you provide the best possible products.

That's how good business works and trust builds up.

So now let's move on to the revenue maximisation.

So I'm going to start with pre-bid, um, uh, for those of you who don't know, and you're in the advertising or publishing or add text space, you've probably been sort of under a rock for a while, um, pre-bid is a pretty big open source project now that's got really big momentum behind it, um, and we fully support uh, pre-bid.

So what we have is a real-time data module that sits within the pre-bid open source.

If you compile that into your environment, then you can start to see more information about the device.

So what we have here is my web browser.

And on the left-hand side, we have a open RTB request without 51Degrees.

Now, this is the one of the 2 bits in this webinar where we're going to be looking at, this isn't code.

This is actually called JSON, but this is sort of the data that flows, not going to go through it all.

The bits you need to see are what's changing between the left-hand side and the right-hand side.

So what you see on the left hand side is what happens with prebid when 51Degrees isn't present, when no enrichment is taking place.

And on the right hand side, you can see the additional field.

So pretty much, everything's the same up until we get to the device extension object, and 51Degrees is adding some information, so FOD, we like to consume as few characters as possible.

Then we've got device ID, so this is the combination of the hardware, the operating system, the browser, and the Corolla, and then TPC is 3rd party cookie, so one means that we've checked and the device definitely supports 3rd party cookies.

That's what I would expect.

Then we have some information that's common across the two, as this is something called structured user agents.

But actually on its own, the information's not too useful.

What does not a brand mean 99, for example, not really very helpful?

We've got a lot to say about that, but that's not for this webinar.

So what we do, through those refined ingredients that I talked about earlier, is turn that into useful information.

So I'm on a desktop computer.

Um, that's device type 2.

We don't know much about it other than it's Windows 10, which is correct.

We do have the IP address, so we've reflected that back.

And then we've got some information related to the location that you can see there.

So that's just it with my browser, with my browser.

What happens if we want 51 did.

So in order to put those identifiers in, some sort of dialogue would have been presented and the user could have a choice over what type of marketing they want.

So I'm going to choose personalised marketing, click this browser again.

The auction's gonna run.

It's running a real auction, not connected to a back end, but it's a real pre-bid auction.

And I now get some extra fields back.

I get my extended IDs, and they've been inserted by 51Degrees.com.

We use 51 d.es just as a shorter domain name.

We'd like to save characters, and then you can actually see these IDs here.

We've got two.

There's one that is scoped to the license key, so it's scoped just to our customer, and the other is global, so it would effectively be the same probabilistic identifier within the overall data structure.

So we can click on the link there, in the pre-bid demo, and we can actually unpack that did.

So we can see that it is valid.

We can see it was created at midday, which is when we recorded this section of the webinar.

Uh, we can uh, see that it has been flagged for these purposes.

So personalised marketing, standard marketing, and non-marketing.

We can see that it was created by our license 19,926, so that's a ID within 51Degrees, so we can tie it back to who we generated it for.

So, remember, I said about know your customer and the insurance and that kind of thing.

That's how we can tie it back.

So if we find these things turning up in the wrong place.

We know who we gave it to.

And then we actually see the probabilistic identifier.

So that's this base 64 string, all these bytes here.

So it's actually this bit that you would compare, if you actually wanted to do a comparison.

And then we have the cryptographic signature at the end, which can be applied back to 51Degrees.

Um, So, uh, what I'll do now is I'm going to change to standard marketing and I don't know, let's be a Samsung Galaxy S 24.

So, the auctions rerun with this new data.

And I've got my device type.

So device types changed.

We can see, uh, we've got a galaxy, we can see the hardware, uh, version is S 24 Ultra, it's running Android 16.

Uh, we've got that device ID.

Um, the IP address has uh, changed.

I think that's because I'm randomising it at the moment.

Yeah, so it's a random my peer address.

Um, and I've got those dids again, so I can click on them, and I can inspect that, and what I'm expecting to see is, there we go.

So personalisation is not allowed.

So only standardised marketing or non-marketing for, say, fraud or frequency capping use cases can be used.

Same license ID. As before.

So, there's a lot of technology behind this. Happy to dive into that on a separate webinar if there's demand.

But this gives you a quick sort of introduction into what's happening within the 51Degrees pre-bud world.

Right.

So, now let's have a look at the business implications of all of this.

So, if you think of an auction or buying a vehicle, you've got certain common information that you would put out there about it.

It's mileage, secondhand vehicle I'm talking about, you know, what type it is, et cetera.

Um, that's pretty much the same.

So the URL, time of day, those kind of things, that's what you saw in those open RTP messages earlier.

But you can choose what additional information you want to show, and that was what I was showing you in that open RTB message.

So you can choose to add one lady owner, a recent service, but you might also, because we're just providing that refined ingredient for you, might be showing things that maybe some people don't value.

So like bald tyres, but can belts broken, et cetera.

So, buyers and sellers value information differently.

And as a publisher, what you choose to provide, or as an advertiser, what you choose to accept, are business choices that you make.

They are not for us to make.

We provide those refined ingredients.

So what we see when all these dimensions replied into advertising is really 3 dimensions.

The device or characteristics of the device, the information we're providing back, the publisher and the bidder.

And those three dimensions basically affect positive and negative outcomes.

Now, there's so many variables, it's impossible to sort of use scientific AB analysis, to determine what the impact is, because 51Degrees is just one dimension, and those other dimensions vary all the time.

So what we see if we take this cube over all of those different dimensions is there are positives and negatives, some publishers and bidders and devices show big improvements, sometimes.

Some don't.

So, for example, if you're a publisher, and you know the device as an iPhone, we've told you the iPhone, we've told you the iPhone model, and then you put that into the open RTB request, and raise the floor price, chances are you're going to make more money because that device is one that advertisers are prepared to pay a little bit more for.

Likewise, if you know it's definitely not an iPhone.

It's definitely a desktop computer, for example, um, and a particular bidder doesn't have the sophistication to work out that it's, um, a desktop and they think it's an iPhone, possibly because it just has iPhone and its user agent string.

A lot of fraud or a lot of spoof devices, et cetera, do that.

Then telling them that it's not an iPhone when they think it is an iPhone, may not be a smart thing to do.

So, that's where you come in, or your SSP comes in, if you're a publisher, or your DSP, if you're an advertiser, as to how you want to handle this information asymmetry.

So what we see is different compatents of the cue, between these different aspects, have positive and negative outcomes.

And that's really what you need to experiment with with these ingredients.

We provide the refined ingredients.

It's up to you to produce the finished goods and see how you use them.

So what we can tell you, from the publisher perspective, there's a lot more we can't tell you, and some of our customers are really making significant gains in this area, but it's their secret source.

What we do know is that the consent unconsented traffic, so reject all, this additional information is always positive.

It gives the buyer more information to work with.

And then raising floor prices, for proven high value devices, like new iPhones, for example, is almost always universally a good idea.

For advertisers, you can start to use your definition of what is CTV.

I think a lot of advertisers would not say a little video playing on a desktop computer, is CTV, if it's sitting there in the bottom right hand corner, compared to a big screen on a wall with 6 people sitting around it.

So, it can start to address those information asymmetries that exist in business, and you can get that edge with 51Degrees.

So, we've been through the use cases, we've been through the raw ingredients, we're going to finish with the process of migrating.

So, what I'm doing here is starting with the demo of Maximind, a popular IP intelligence provider, we're taking their example, a slightly reworked version of their example that they provide online as open source, and we're going through the changes that are needed to make it work for 51Degrees.

So, firstly, we change the import, so we're bringing in some the 51Degrees components.

That's step one.

Then we need to change the data file that we're using from a maximine data file to a 51Degrees data file.

Then we're gonna switch over and use the 51Degrees pipeline, so that we've actually instantiating the 51Degrees solution.

We could also add device detection into that pipeline, so you can run the 2 together, but that's outside the scope of this little coding demo.

And then we are unpacking the property so that we can show you the results.

We're gonna then run it again with the same IP addresses, and you can see different results coming back from 51Degrees.

Pretty much a good match on countries and things like that.

We've added the confidence value from 51Degrees in there as well. That I talked about that I talked about earlier.

So it really is that simple, very quick, five-step process.

I know we've got a broad audience for this webinar.

So that one's there for the coders.

If your coders come back and say, oh, this is hard or we don't have engineering time.

Then please direct them back to this recording.

It will be up on the website in the next couple of days.

Um, and we'll, sorry, I'll be adding more information about migration uh, in the coming weeks.

Pretty similar for, you know, other solution providers, but, you know, maximum mind has a legacy following that's quite popular.

So switching out that refined ingredient to 51Degrees gives you a lot of benefits.

So that's our whistle stop tour of the different ingredients and how you can use them in advertising, publishing ad tech.

We've got time for questions.

Alex is going to join me, and we'll take your questions.

All right, Alex, thanks for joining me.

That's all right.

Let's get to the questions, then.

Yeah.

So, the first question on here is, as follows. 51Degrees has two entries listed under IP location services on the ad com spec, whereas others only have one.

Can you explain the difference and the reason for two entries as opposed to one?

Oh, someone was paying attention, yes.

So we've been live in this space for some time.

We've got quite a few deployments going out in the sort of early adopter stage in the 1st 6 months.

So, in the Adcom specification, the question refers to, there are different providers of IP location services, and we have the entries 511 and 512.

So 2 entries for one vendor.

The reason for that is those 3 location confidences I touched on in the webinar earlier.

So, if the location confidence comes back high, then use 511 as the vendor, so that anyone downstream can go, OK, this came from 51Degrees, and they have high confidence in the information that's being provided.

If it's medium confidence, then it comes back as it should be 512.

So the pre-bid JS that I showed earlier.

So that's our real-time data module for pre-bid, that will automatically handle the setting of those 2 values.

If it comes back low confidence, then that means that you shouldn't send the location information.

So that's why there's no 3rd entry for 51Degrees.

And that's because the ad com and open RTB specification don't provide degrees of confidence.

It's just not there in the specification.

So, um, it was agreed, uh, with IB Tech lab that we could use two vendor IDs, uh, to signal the two differences.

Great, very clear.

The next question is, what happens if somebody else in the chain breaks the standard agreement?

Oh, yeah, so we're referring there to the 51D ID.

So we have these standard terms that are used in everyone's contract.

So I think what you're referring to there, or the question is referring to is the, we enter into a contract, would say, a publisher or an SSP.

I explained earlier about the know your customer, know your business, and then the insurance that we require to cover the indemnities for breach.

So a breach could be, probably the most likely breach is going to be using the data in a way that's contrary to the contract.

So an example might be identity linking.

So that's absolutely prohibited in this common contract, I think, most participants in the ecosystem would say, re identifying is not allowed, That's not a reasonable expectation that anyone have.

It's not in any contract.

It's really prohibited by most best practice.

Um, so, uh, let's say it's then passed on to some, some other party that then breaches that agreement by re identifying.

You know?

Obviously the crime has to come to light.

So it's like speeding.

You don't prosecute all instances of speeding, because you don't have enough speed cameras to find all the instances of speeding.

But if it comes to light, we've got that cryptographic signature that I showed earlier.

So what that can be used for is to say, OK, this definitely came from us, so we can go, yep, it definitely came from us.

We were the original source of this.

Where did it turn up?

Yeah, how did it get to where it turned up?

Who did we provide it to?

I mentioned the license ID, that number in the ID earlier.

So where did that turn up?

And if it turns up, um, obviously, with our customer, you know, we know our customer, and it was them who did the real identification, then that's very straightforward, but I think the question is referring to it being many hops down the line, um, and what we would then need to do is work with our customer and say, we gave you this ID.

Who did you give it to?

And so on and so forth until it's been found.

So, um, I think, you know, there might be a lot behind that, or, more certainly, is a lot behind that question, um, that is not perfect, there is no perfect, um, you know, in detecting speeders or, uh, in this particular world, um, but, uh, we believe it is way better, way better than what happens today.

Yeah, ways you don't have the possibility of even finding the bad actors in many cases.

So that's what we mean by improving transparency.

And just, I think this is a feature of this approach that is really, really important for everyone to understand before they go, Oh, you don't have encryption, so how can it...

How can it be any good?

Which, like, encryption still requires a contract.

When you unencrypt something, you still require a contract, and those contracts are opaque, they're not open, and that's why I think this is a net benefit. To absolutely everyone.

A related question, I think, here, which is about insurance.

And that is that with the identifier, Who needs to hold insurance and who needs to agree to the terms in that chain?

Yeah, so the contract is a standard contract.

That contract doesn't require insurance?

And it doesn't necessarily require indemnities to the other party to adhere. To that.

That is a business choice of each individual participant.

So, what we're doing is 51Degrees, as the initiator of the contract is requiring, uh, that, that insurance for the, for the indemnities that we put in our contract.

So our contract, with our customer, will say, were to the effect of, you will follow these terms, you will indemnify us for any breach of these terms by you, um, and, uh, we require insurance to cover that if they don't, uh, I mean, there are other things, but basically their insurances, the, is the, is generally what's needed if they already have.

Um, so a lot of our customers will already have um, data protection insurance, um, already in place.

So that might be enough.

It's just a case of seeing the, you know, seeing that insurance certificate, but it needs to be backed up, basically, we can't have a situation where someone doesn't have the means to hold up the indemnity and the adherence to the terms document.

Okay.

Um, Next question.

Do you have any tips for tracking invalid traffic?

Well, I showed some of them in the recipes piece earlier.

So I, I, we covered a lot earlier.

The recipe analogy, the refined ingredients, you know, we're providing those ingredients up to you, how you deploy them.

I listed some of the examples earlier.

So frequency of requests.

You can use the 51 did.

If you go back to the pre-bid page, you can set non-marketing, for example, and that won't be bound by these contractual terms, because we don't need a contract for that particular uses.

It's just between us. Customer.

That ID can't be passed on, if it's got that flag set issue.

It can't be passed on, that's prohibited, so it's just internal use.

So you could use that to look at frequency of requests, for example, high frequency might be an automated bot.

You've then got the age of the browser.

So if the software is or the browser is, say, 3 months old, not the latest version.

Why is that?

Most people's browser versions update.

All the time, you then got the pairing.

So maybe it's a very unlikely pairing of browser operating system.

We often see iPhone, you know, popping up, but also being confused with Android, for example. Some of the less well thought through scrapers, et cetera, do that.

You've got human probability.

And then around location, you've got that location confidence.

And then, of course, those are the crawler properties and things like that.

So, those are all dimensions that you, or, you know, ingredients effectively that you can use, but ultimately, it's not our job to just go, oh, this is invalid traffic or not invalid traffic.

There are other solutions out there to do that.

Of course, use them if you want to use them.

What we're doing is providing those refined ingredients, so you can easily mix them together and maybe come up with something better.

Yeah, maybe actually look within your environment, at the history of what signals have led to, you know, advertisers not getting a good experience, for example. And then downgrade those.

So maybe it's not just invalid traffic.

Things we don't want to touch, because historically, they've not worked out well, might be a better definition.

Great.

There's a question here about the know your business checks.

If you just remind us what that means.

And the question really is what form do the checks take?

Yeah, so, um, there's 2 different ways of subscribing to us basically.

There's the online. And there's the offline bespoke agreement effectively.

So in the bespoke agreement, We need proof that you are, who you say they are.

So that might be notarised documents, for example, and if you're not in the same country, it's something that basically says, yeah, we absolutely know that you are this legal entity.

Often, it could be passport, you know, copies of passports or something like that.

For the self-service where you're coming in online, then we use a third party.

Same as I think a lot of people, particularly in the UK, will be used to the service that comes up and solicitors use it, banks use it.

You have to hold the phone up and turn to the left, turn to the right, say a few words, things like that.

So that's proof of identity.

So we know the person is who they say they are and then they match the records at company's house and things like that.

So it's a combination of those two.

We're really not reinventing the wheel there.

But we believe, you know, in integrity.

We believe in quality, and therefore, that's an important kind of first step.

Otherwise, those other things don't bite.

Yeah?

This is about building a partnership, and we have to know who our customers, for this particular piece of data.

Great.

Okay, I've got a couple more questions before we wrap up.

The 1st is for really, on behalf of anyone watching, where can they find out more about migration?

Yep.

Can we share my laptop screen?

Yeah.

You making the magic happen.

Cool #um.

So #ah let me go back to the beginning so we've got the developers menu you come to the 51° website you've then got latest documentation.

Um, you've got device detection, um, and uh, you've then got migration guides, for example, for device detection, um, and we're just adding them in for uh, IP intelligence, uh, in the next couple of weeks, uh, they're coming in.

So there's two parts to the migration.

One is, as I showed you at the end, I know, we've got quite a mixed group on this particular webinar.

So for the coders, you'll be going, okay, that doesn't look too difficult.

If you're not a coder, um, tell your coders, look at this video, um, it really is very, very straightforward.

We will provide more resources as needed, if necessary.

Clearly, it's in our interest to help with migration.

But it's really not very hard.

That's the technical aspect of migration, and then you've got the data mapping.

So that's where you would see tables like this, which show how different properties are mapped from one vendor's vocabulary to another, and how to handle that.

So, an example would be location confidence.

You might have seen that at the very end, of the brief piece of code that I was showing, where we don't, we have location confident, so we showed that in the results, whereas Max Mind just had the accuracy radius, which we have an approximation of, but we don't treat the world as circles, we treat them as irregular polygons, as you saw, earlier geographic service areas.

So there'll be some differences.

So you might have a reasonable one to one mapping, and then there'll be new features from 51Degrees you can overlay later, that it really is very straightforward.

And particularly now with the resources we provide, and if you happen to, like, using an AI to help, then because everything we do is open source, it knows pretty much everything about how everything works.

So, it can do a pretty good job for you, whereas if you're using a closed, you know, service like a digital element, or a CNG mobile, you're obviously going to have to feed the AI, that code, and let it work it out, it's not going to have the benefit of over a decade of experience, that, um, AI sort of software developer helpers going to have with a 51Degrees product.

One of the questions has just popped in before, which I will squeeze in before I ask my final questions, and that is, does 51Degrees support go and pre-bid server?

Yeah, yeah.

So, uh, we still sharing much?

I can go back to your screen.

Cool.

So you go, 51Degrees, pre-bid, server, let's see what it comes back with.

There we go.

So this is our pre-bid page, uh, for pre-bid, um, and we have everything in pre-bid go.

This is being updated to support all the features I showed you earlier with pre-bid JS.

Although, you know, we're delighted a lot of our customers have handrolled there, their integrations already.

Prepids a bit like that, you kind of have the main open source, and then you have bits on the side where people create their own, you know, internal secret source.

Those ingredients I talked about earlier.

I mean, Prebids are pretty big, a big ingredient, but it is just an ingredient.

It's not a finished product on its own and that's the beauty of open sources, how you assemble those ingredients, which is what's giving you the competitive advantage.

And we also have the integration with Java as well, which you can see off the same page.

Great.

So I'm going to wrap up, I think, with...

I mean, it's 3 questions in one really, but they're all related, and that is, how do you sign up?

Can you try before you buy?

Is there a trial, is there a trial that you can use for to try these things out?

And what does all this cost?

Cool.

So the sign up, any of the tested pages, in fact, let me start.

Yeah, the test of pages.

So I don't know how to take IP testing.

You see free deployment in minutes with our cloud service and then down here you've got a number of links.

If I start with the cloud service.

Basically, you click on those links, it'll pre-populate some of the properties for you.

Um, basically, you've got those properties there.

Some of them might be paid, some of them might not.

You then move on to the 2nd step of the 3 steps in our configurator, so you can choose free, if you want, but I have selected some non-free properties.

So let me just get rid of them for a minute.

I'll just take country.

Um, so I want my free property, um, and then I can go through.

We do suggest you get service updates.

You can put your domains in, and then you can say, yep, I'm ready to go.

You've then got your resource key, which underpins those things, and you can see client side deployment.

If you're using device detection, do you pay attention to these permission policies, it's a bit of chicanery that Google added, which no one really wants.

It's pretty useless or pretty unnecessary.

But anyway, if Google want to do it, you kind of have to do it.

So, um, you get your JSON, and there you go, you're getting country back.

It really was that simple.

If you come back here, you can then also choose to enter a license key.

If you add that, then the license key that you purchase through a different process, which I'll show you in a minute, can be used, and that's what activates those paid properties.

So those are available back at 51. 51Degrees, so you come to pricing, so free, I went went through earlier, if you want some extra properties, you can click on these links, so you can see a table of the properties that come with the essentials, subscription, monthly, €49 or annually, if you want.

And then you can buy these, it's a very simple process, as you can imagine, you come in here, tells you what's going on.

Um, and you check out, just like normal, we use a third party maxio, um, also in a charger fight to handle, um, the subscription.

So you've got your free to try, if you want cloud, and then you've got essentials for those extra ones.

We don't do trials for the essential properties.

I guess it's reasonably basic if you are not prepared to commit €49 to do a one month, then the admin cost for us.

It's just the economics don't add up.

If you want on premise, then you need to come here here to bespoke, choose the product that you're interested in, and then one of our lovely team will be back in touch with you to find out more about what you're looking to achieve.

We want to make sure that our ingredients are properly understood.

We're not quite at a point where all these things are just like flour or eggs or milk in the supermarket.

They do need a little bit of explanation.

So we want to build partnerships.

So, it's really important, particularly from premise, that we understand what you're trying to achieve.

There's some very demanding compute environment.

It's very specific timings that are needed.

So, a little bit of handholding is needed there.

And as you can imagine, the price is a little bit more for the on premise options than it is for the cloud.

Great.

Well, that's it.

I think you've answered all the questions.

Unless you have anything else, I think we are done.

Well, just to thank you and the team at 51Degrees, and this is the last of our planned series of webinars.

So thank you for production, and being question master, and the team back at 51Degrees for all their help and support.

We can take a little break, why we just assess what we want to talk about next, effectively.

You can imagine that there's certain hot topics coming up, like the content protection, as that rolls out.

That could be quite significant, as we end up changing the value. For publisher information, and it starts to be valued in the same way as chips and energy.

For example, we're delighted to be playing a role, particularly with small publishers, actually.

All the free stuff that I was showing earlier, gives that basic bid up to enforcement that any publisher would need.

So there's a lot of free tools there.

They can just deploy it and we can just keep it running for them.

We're very happy to do that without the vibrant open web ecosystem.

Then I think society's going to be a lot poorer for it, and we want to play a role in protecting that.

And then, of course, there's the rollerout of the ad tech monetisation side of things that advertise a publisher piece I talked about earlier with the queue.

I think there's going to be more to say on that and perhaps we can do that with some of our chefs, effectively, who are building the recipes.

They might be happy to sit with us sincere at least some of what they're doing.

So that might be coming up soon.

And then finally, thank you for watching.

Thank you for your attention.

Thank you for your interest.

Please get in touch or keep in touch.

We only work when we understand what customers want, what's causing their problems, that helps us create better refined ingredients.

Wonderful, thanks very much.

Thank you.