What is Programmatic Advertising?
Programmatic advertising campaigns allow marketers to buy native ads on a variety of publisher's sites while using smart algorithms to ensure that they are targeting the right viewers at the right times all while remaining within budget thanks to a bidding system that allows them to compete for impressions with other advertisers. In short these campaigns give the precision and quality of native advertising (like banner ads) while at the same time allowing for the control, adaptability and precision that you would get with a PPC campaign. The net result is ads appearing on the sites of carefully selected publishers, but only when and if they are likely to yield the best results. Programmatic advertising uses a complex algorithm in order to identify your business’ ideal customer, and then to figure out where they are likely to appear on the web. It will then show the ads in those places and then utilize a learning algorithm to Instead of going from one publisher to another discussing rates, you allow a “bot” to do all the work for you. More importantly though, this means that you won’t waste money on an ad slot that no one looks at. Your ads will be chosen and honed by a smart algorithm, and the result will be that they get higher CTRs from the right targeted customers!
RTB
This is how programmatic buying is different from either PPC or just buying a banner ad on your favourite news site. Giving you further control meanwhile is the option to use RTB or “Real Time Bidding.” What RTB essentially means is that you are automatically entering into a bidding war each time a page loads based on your pre-defined budget. These bidding wars allow your ads to compete with other advertisers around a large selection of different websites that all cater to the specific demographics and context that you have chosen. In other words, you will define that you want to target sports sites catering to an audience of males in their 20s-40s, and from there your ads will appear across a selection of those websites (which you can still curate in some cases if you wish) based on the outcome of each little bid. This allows you to target your audience across a range of different sites and to avoid paying too much to do so). On the other hand, direct buying is essentially a bulk order of impressions from a specific website/website(s) such as ESPN.
You can still filter your impressions by a range of factors such as location or browser, but essentially you will be targeting a specific site and securing a place on that precise spot. Direct buying is effectively a little more like advertising with a banner ad, while RTB is a little closer to the PPC model here you bid for spots across a range of websites (while still having a little more control). Deciding what will work best for you will depend on a number of factors. For instance, your budget is going to come into play as direct buying tends to be more expensive (because lower CPIs won't have as many opportunities to appear). It's also worth considering that RTB gives you more flexibility, more data and more control – you can identify which sites are working best for you, at which times and for which viewers, and then you can further tailor your approach accordingly. On the other hand though, if you bid too low using RTB, then you risk your ads not appearing at all. This is in direct contrast to direct buys which essentially “guarantee” you your ad spot and that you will eventually be certain to get that number of impressions.
This is useful for a company with more concrete goals and a fixed timescale. Likewise, with direct buying you will be able to more tightly control where your advertising appears and forge a closer relationship between your brand and that of the publisher. So different approaches will work for different marketers and for different purposes. Your job is to to decide which works best for you, and the best way to do that may well be to dip your toe in and try them out for yourself.
How to be More Successful With Programmatic Buying
Programmatic advertising has very quickly risen to prominence in the online marketing industry and is an important new tool for any business that wants to reach a bigger audience. That said though, even the best programmatic tools are only as effective as the people controlling them, so before you get ahead of yourself consider these four crucial tips to ensure your success.
Don't Forget the Creative Component No matter how well targeted your ads are or how smart your campaign is in terms of exposure, if you haven't invested in the creative aspect then those ads are going to fall flat. Design your ads well and test to see which designs work best. Likewise, think about your brand identity and how you can strengthen this even through ads that don't get clicked. Choose a programmatic partner who can help you with this aspect.
Consider Audience and Context Before choosing which publishers to work with you need to look at the audience they are attracting as well as the context. The ideal partner will be a site that writes about topics related to yours and targets the precise same demographic you are. In some cases though you won't be able to find both, and thus you will need to select publishers that provide the best balance. And don't be tempted to ignore context in these situations, because studies show that the same person is far more likely to click on an ad for golf clubs when they're on a golfing website compared with a news site. More to the point, someone who wants a wedding dress is only going to want that wedding dress during a particular time in their life and this really demonstrates the importance of context.
Be Willing to Spend to Begin With The beauty of programmatic advertising is that you can directly manage your spending in real time to ensure you get the maximum exposure whatever your budget. However, it's highly recommended that you start out with a higher spending-limit than you intend to continue with, as this will enable you to more quickly identify what works and what doesn't. Remember, you want more data and that means you want more clicks. As you tweak your campaign in response to stats and ROI you will get closer and closer to the optimum set-up, but if you don't spend the money up front then you won't be able to tell if your campaign is working because you won't win enough bids. Spend a little up front and you'll save money in the long run once you settle into your groove.
Make Sure Your Ads Fit In The downside of any automated advertising campaign is that you can lose that “personal touch” – the benefits that come from working with a publisher to develop an advertising campaign that matches the tone and appearance of their overall site and that they will help to promote throughout their content. Unfortunately, native advertising is difficult to scale which is why so many automatic platforms are popular. To be successful with programmatic advertising this is something you need to consider. Your ad is targeting the right people, on the right devices in the right contexts. But is it the right fit for the job? This ad needs to look like a native ad in the same way that a banner ad would. Creating ads that will blend into a number of “environments” is one way to do this. Another is to choose a tool that allows you to select the brands you want to work with, and then to choose those that are already closely aligned with your own goals and style.
The key to success on the web is not just to gain traffic but also to control that traffic. What does that mean? It means that you need to know how to decide which of your visitors you want to talk to at any time. It means that you need to understand your visitors and to know what they’re thinking, what their moods are and what they’re interested in at any given time. And it means you need to know how and when to strike when it comes to selling products or encouraging people to sign up to your mailing list. This is a theme we’ve seen come up time and again with machine learning and AI. And email marketing is no different. You can do all this by building a mailing list and then segmenting that list. First, let’s go over the basics of email marketing again for those that aren’t familiar.
Email marketing is of course the process of marketing via email. In other words, this means you’re going to be building a big list – a collection of emails – and you’ll do this by asking visitor to your site to share their contact details when they land on your home page. This in turn requires an autoresponder. An autoresponder is a tool that you use to create email forms and then to manage all of the contacts on your list. You can use the form somewhere on your page to let people input their details and you’ll use the autoresponder to actually send all your emails. It should be immediately apparent what the value of this is. Sending all your emails manually using Gmail or another web client is not easy and would likely result in many not getting delivered. You’d have to send lots of different emails for longer lists and you’d need to manually manage any requests to subscribe or unsubscribe. An autoresponder manages all that for you, so you just need to write one email and then click “send.”
The other benefit of an autoresponder though is that it can collect data for you and use that information to do a range of different things. For example, an autoresponder can show you the percentage of subscribers who actually open your emails. If your email subject headings aren’t successfully encouraging people to read, then you can identify this problem and work on a solution. Suddenly, we’re using a data-driven approach and machine learning again! You can then see all the visitors who did read a given message in one place. Or choose to see all the ones that didn’t. You can see the open rate for different individual visitors and you organize your list by different factors.
That’s another handy thing about using an autoresponder: it will allow you to grab more information using the form embedded on your page and that information can then be used to group your visitors. Want to just message the men? Go for it. Want to just message the people over 30? You can do that too. Or how about having multiple different mailing lists for different brands, or even for different products? All of this can be accomplished using just a single autoresponder. And of course, this kind of control and automation opens up all sorts of possibilities when it comes to AI and machine learning for marketers.
Lead Warmth and Email Segmentation
The true power of all this information comes from being able to use it that data in order to pick and choose who your messages go to. For instance, you can decide that you want to send an email only to people who fall into particular categories. What we’re interested in to begin with is sending emails based on engagement, retention and lead warmth. A lead is anyone who has shown some kind of interest in buying from you. That means that anyone who has signed up to your mailing list can be considered a lead because they have demonstrated an interest simply by doing this. But at the same time, a lead is also anyone who visits your site, or who takes your card. This is a ‘cold’ lead, whereas someone who actually gives you their contact details is a ‘warm’ lead.
Leads get warmer the more interest they show in what you do and what you’re selling. And the warmer a lead is, the more likely they are to buy from you. In fact, this is the true and most useful purpose of having a mailing list to begin with: it allows you to take your ice cold leads and turn them into warm leads and then paying customers. I always liken this to asking for someone’s phone number. If you were to just walk up to someone in a club and ask for their number, they’d likely just tell you to go away. Why would they give you their number when they know nothing about you and have shown no interest in you? First, you need to chat to them and let them get to know you. If they look at you and smile, they’re a cold lead. If they respond to your witty banter and tell you their name, they’re a warm lead. If they’ve kissed you or let you buy them a drink, they’re a hot lead. And once they’re hot, you can ask for their number.
This is all about timing. Time this wrong and they’re not going to give you their number because you haven’t laid the ground work! The exact same thing is true with internet marketing. If someone visits your site and you tell them right away to buy your product, they won’t. Why would they? You haven’t given them any reason to trust you. You haven’t told them anything about you. They don’t know much about the product. Ask them to hand over their email after a few blog posts though and you can gently start to increase engagement. This is then when you wow them with all your information and all your knowledge. You entertain them a little and you let them get to know you. If they don’t open your emails, that’s the equivalent of giving you the cold shoulder. That’s like the girl or guy in the club that isn’t laughing at your jokes and keeps looking away. If you try and sell to them now, you become spam. And you get deleted. And they never return to your website. But if they open your emails, you know you’re in with a shot.
That means you can then send them some more information about your products and get them excited for your product launch. If during that they still keep opening your emails, then you know you’ve got an even better chance of success. If you now try and sell to them, there’s a much better chance they’ll buy from you. Using email segmentation you can do exactly that: you can see which of your visitors are actually opening your emails, are actually clicking your links and are scrolling down to the bottom. In fact, using cookies it is even possible to see which of those visitors has been to your website and looked at your products. You can see who has hovered on your products and been tempted to click buy.
Email Segmentation Combined With Machine Learning
If you’ve been paying close attention, then you might already have figured out where this is going. Remember when we discussed big data, we said that you could use predictive analytics in order to better rate leads? This is where things could get really interested for the autoresponders of the future. Imagine if your autoresponder not only segmented your audience and looked at the open rates and engagement, but if it could look for trends across huge data sets. In other words, what if your autoresponder recorded every single person who bought from you, and assess what actions they usually took first.
This would allow that machine learning algorithm to better spot when a user is behaving like someone ready to buy, and to send them a message tailored to encourage that purchase! This could be combined with smart recommendations in order to improve their likelihood of buying even more. This is already being used by some big brands, so it’s only a matter of time until more of us get access to this same amount of precision.
Tips Marketers Can Follow Now
For any of that to work, you also need to make sure that people are signing up to your mailing list in the first place. Again, this will help us to prepare for the AI-powered future. There are a few ways you can encourage this. Firstly, make sure that you show your mailing list wherever you can. At the very least, that should mean that your mailing list shown at the bottom of your posts. At the same time though, you can also place this in the side bars so that your list is visible on every page of your site. Another tip is to make sure that you draw attention to it. A mistake a lot of people make is to create their mailing list and then just ‘hope’ that people see it.
Far more effective is to occasionally tell people about it and to explain in your posts why it’s a good opportunity and why people should be excited to sign up. Here’s the thing though: you should always be honest. The aim of a mailing list is not to grow it as much as you possibly can. Instead, the aim is to grow it as much as you can with only highly targeted visitors. If your visitors have no interest in what you’re offering through your list, then you will just frustrate them and effectively be spamming.
Chatbots are an increasingly popular tool for marketers, business owners, and webmasters. So what is a chatbot? Essentially, a chatbot is a miniscule AI that will normally live on a website, and which will then be able to answer questions and engage in basic conversation. Very often, chatbots are used in customer service. This way, a website can answer commonly asked questions and significantly lighten the load on its customer service team. The company can provide the support it wants to for its buyers, without spending a huge amount on extra members of staff and call centers. But a chatbot can be much more than just customer support. Chatbots are just as effective in marketing, and can be extremely effective at increasing sales and profits. Chatbots are particularly effective at kickstarting a sales process, by welcoming visitors to a website and asking them what they’re looking for.
Instead of relying on UX to try and guide the visitor to the right part of the page, a chatbot can instead simply ask what they are hoping to buy, and then take them to that page. What’s more is that it can provide useful recommendations (perhaps informed by previous buying history) and it can reduce any concerns that users might have. Chatbots can even get information from customers, by asking them their budget, or what it is they’re looking for. Even if they don’t buy, you now understand their intent and you can use this information to hone your marketing strategy further. Some pundits argue that in the next few years, 85% of business might be done through chatbots! So how can you get started? 80% of businesses say they want chatbots in place by 2020!
Facebook Chatbots
One option is to invest in a Facebook chatbot. There are plenty of sites and services online that will set these up for you. Facebook Messenger is a new frontier for a lot of small businesses. While it has crept under the radar for many marketers, the numbers speak for themselves. Facebook Messenger is currently used by over 1.2 billion people! That’s 11% of the entire human population. And what makes Messenger even more powerful, is that it can be embedded into your website. Over 20 million pages use messaging and that number is growing. This provides a very simple and easy way to communicate with your visitors, to answer their queries and to help convert traffic into sales. But you can’t be available 24/7 to cater to all your visitors’ needs. And this is where a chatbot comes in. This is a basic AI that can tend to your customers and help to provide a more personal experience while answering basic questions.
This is a hugely powerful tool for businesses as it means you’ll no longer lose visitors that struggled to find their way around the website or to get the information they need. Imagine if customers could order food by just sending a Facebook message and then answering a few automated questions? Or what if a business could access your expert legal advice without ever needing to speak to you in person? All of this is possible in the near future. Facebook chatbots can even be proactive by sending messages to potential customers. You need to be extremely careful with this, seeing as it can be seen as spam. But if you have an automated system that’s able to reach out to potential buyers at just the right time with a carefully crafted message, this can be huge for business!
Other Types of Chatbot
Of course, not all chatbots are Facebook chatbots. There are many other ways to implement chatbots into a website, from creating them from scratch using home-made software, to using them to respond to emails or SMS.
Relational databases consist of numerous tables like you see in Excel, with columns and rows. So if you had a list of visitors to your website, you might fill out their data across rows, such as name, age, contact details, etc. Pull out any given visitor, and it will bring their details up so that you’re ready to call them and market to them. SQL then allows you to do things like creating whole new tables, or inserting new rows, columns, or cells. You can do this with simple commands like “CREATE TABLE” and “INSERT INTO.” To make a new database, you first need to use a command to make it, and from there you can then begin inserting tables like so:
SELECT FIRSTNAME FROM CLIENTS
WHERE AGE > 23;
GROUP BY is a command that lets you group results according to certain conditions. CURSORS let us move through sets of data and make changes. While this all might seem quite simple, combined with huge amounts of data, these simple commands can yield fascinating results and be extremely useful in informing future decisions. This is essentially how machine learning works, and if ever you want to work as a data scientist to employ big data solutions or machine learning applications, this is what you will need to know.
Over the course of this ebook, we’ve examined a large number of different types of AI and machine learning in the context of digital marketing. The objective of all of this has been to help you better prepare for the future. You know that you should start collecting as much data as possible right now, that you need to add schema markups to your site, that you should be using LSI, that you could benefit from a chatbot… You even know a little bit about SQL, just in case you ever decide to get involved behind the scenes! But in all likelihood, all of this is going to change a lot more before it enters the scene – and we can’t really anticipate just how its impact will be felt. There are huge waves making their way through internet marketing, and the power of AI can’t really be understated. Imagine for instance, what will happen once AI that can write high quality content becomes commonplace and commercially available.
This technology already exists – AI that can write nearly as well as a human – but when it is allowed to cut loose on the web, it could potentially flood the internet with enough new content to double or triple its size in a matter of days! How will we know what’s written by a human and what’s not? What about when AI can create realistic looking images? We’ve already seen the power of deep fakes: how will we know what’s real and what’s not? We can’t really prepare for these scenarios because we don’t know how they will play out. So for now, it is best to focus on what we do know. And in particular, as a marketer, that means focussing on Google’s move to natural language processing and more human-like interactions. It also means that Google is going to keep getting smarter. Google used to look for keyword matches, but now it actually understands the meaning of a website and can use many more metrics in order to understand whether it is high or low quality, and whether it is delivering on what it claims to do.
In fact, Google likely has the potential to become the most powerful AI on the planet, owing to the huge amount of information it has access to, and the huge resources that the company pour into it. This makes it increasingly difficult to ‘game’ the system or to try and trick Google. The best thing we can do then? Make the best quality content we can. As Google becomes more and more human, writing for Google and writing for the reader will mean essentially the same thing. It’s time we started focussing on great quality content, and on providing real value. The key thing to remember is that Google serves its customers first and foremost. Who are its customers? The users who use it to find information and entertainment. Google wants people to keep using its search, and as such, it needs to make sure it consistently brings up only the most relevant and interesting information.
As long as you are concerned with creating great quality content for your readers, then your goals and Google’s will be aligned. Thus, each time it gets a little bit smarter, that will work to your favour instead of being something you need to worry about. As Google gets smarter, it will find more ways of identifying the best quality content. And so if you’re focussed on delivering that, Google will find more ways of connecting you with your audience. Combine this with more data collection, and a generally more data-driven approach to marketing, and you will be ready for the future of the industry
Posts
No comments:
Post a Comment