For those of us in advertising circles who have seen firsthand how Artificial Intelligence (AI) and machine learning can help us do our jobs better, it is natural to wonder whether AI-enabled computers will ultimately replace humans for critical functions in the industry.
Already, computers are replacing jobs in business across many industries, whether its finding ways to automate the assembly line for manufacturing plants or taking lunch orders at your local McDonald’s. In fact, a report from the White House issued last December forewarns that AI could threaten up to 47 percent of US jobs in the next two decades.
There’s no denying the fact that computers are simply better than humans at some advertising tasks. AI not only automates many of the labor-intensive processes that were previously handled by humans, but it performs them far more effectively due to its ability to analyze large volumes of data and make much faster decisions than any human ever could.
Nowhere is this capability more obvious than in the role of a media buyer. Traditionally, media buyers have had to handle several labor- and data-intensive tasks: understanding the whims of the market and adjusting bid rates accordingly, developing new creative that resonates with targeted audiences, and managing all of the placements where their campaigns run, to name a few.
And yet, today’s media buyers are tasked with managing campaigns that involve greater scale and exponentially more data points than ever before. With the media explosion caused by mobile apps, social networks, and other forms of digital entertainment, many media buyers are now managing a dizzying array of campaigns, placements and media partners compared to the past.
Given their ability to process data much faster than any human ever could, AI-enabled computers have replaced humans in managing many media buying functions -- to great effect. Whereas in the past media buyers have used rules, excel spreadsheets and queries to get the job done, savvy media buyers are now relying on AI tools and algorithms to do those same tasks.
Take targeting, for example. Programmatic advertising has become a nearly $33 billionindustry largely on the back of AI and machine learning. Media buyers find programmatic advertising so valuable because it uses AI to analyze immensely large sets of data about demographics, interests and purchasing preferences in order to determine the best audiences to target for a specific ad campaign. All of this happens in real-time, enabling media buyers to find a level of targeting and efficiency that was impossible before AI came along.
Or look at bid rate optimization. While media buyers can do do their best to keep tabs on what their top competitors are bidding and how rates fluctuate with the times, computers can monitor every company bidding on related search terms and identify fluctuations as they happen -- then make changes in real time. We are already seeing some of the major platforms roll out AI-based bidding optimization, such as with Facebook’s App Event Optimization, which uses machine learning to identify users that are likely to actually engage with an app or perform other valuable actions, not just install it. The solution even recommends how much to bid on those users by predicting their future value. Snapchat, meanwhile, launched Goal-Based Bidding (GBB) late last year.
AI is even beginning to infiltrate the world of ad creative, a realm that many considered the one area that humans, with all of their creativity and imagination, can do better than computers. Automaker Toyota, for instance, used the AI tools of IBM’s Watson to develop a series of customized ad creative for its Rav4 crossover. One example targeted consumers who had demonstrated an interest in both running and luge racing to serve ads encouraging them to try out a gameshow-like activity called ”Win Luge or Draw,” in which they draw pictures of movie scenes for team members to guess before the other team completes a 26.2 mile icy luge course.
All of this might lead you to believe that AI is on the cusp of replacing the media buying role. But we shouldn't be so quick to draw that conclusion. While computers are great at performing manual tasks that involve making quick sense of Big Data, they are no match for human intelligence when it comes to looking at the big picture and synthesizing effective strategies. Advertising is, at its heart, a creative endeavor, and there will always be roles for people who are able to see things in different ways, relate to audiences in earnest, and solve problems that others had given up trying to fix.
In today’s fast-moving landscape, approximately 80 percent of a media buyer’s time goes toward tasks to simply maintain a campaign’s performance, such as bidding, budgeting and creative optimization. The other 20 percent of their time goes to actually growing their campaigns, through strategies such as identifying new channels, analyzing user behaviors, hacking new creative, segmenting CRM databases, and so on.
In practical terms, this pitting of man versus machine represents a false dichotomy. While there are sure to be some jobs lost to computer automation along the way, it is wrong to assume that we are playing a zero-sum game.
By using AI and machine learning to automate a lot of the menial tasks that fall under their job descriptions, media buyers are able to free up their time to focus on other, higher-level objectives. Rather than spending half their day uploading new creative, running A/B tests, and adjusting Insertion Orders, media buyers can let computers handle all of that for them while they work on more strategic goals -- discovering new audiences, for instance, or breaking into new territories, brainstorming new creative, and so on.
Artificial intelligence is coming to the advertising industry whether we’re ready for it or not. In fact, it is already here -- not to steal our jobs, as some have forewarned, but to help us do those jobs better. If advertisers want to do their jobs more creatively and strategically, it’s time we let go of the “man vs. machine” mentality and embraced our new “teammates.”
by Peli Beeri
One of the increasingly popular remote access techniques to grant teleworkers access to internal corporate applications and data is to allow them to log into virtual desktops. While a virtual desktop infrastructure(VDI) can be operated on-premises, cloud-based VDI has plenty of benefits. Cloud-based remote access via VDI is often called desktop as a service(DaaS), and it takes away the upfront cost, buildout and management complexities from internal IT staff and offloads those duties to a cloud services provider. A properly tuned DaaS could be the most effective way to offer internal computing resources to remote workers around the globe.
If you prefer to use more locally deployed, software-based remote access technologies like IPsec or SSL virtual private networks (VPNs), the cloud can still assist. Many IT departments have discovered that moving their authentication mechanisms out of their private data centers and to cloud-based remote access allows for easier management and a more streamlined approach. If yours is like many organizations out there, you likely have some apps and data in the cloud and others in a private data center. Early hybrid cloud designs often left the authentication component in the private side of the network. However, now that most organizations are more comfortable with the security and stability of public cloud services, they have found that moving the end-user management and authentication to the cloud allows for a more centralized management experience for both publicly and privately hosted company resources.
In situations where staffers work out of small branch office or teleworkers work out of their homes, many companies are opting to build a different sort of remote access: a static, site-to-site VPN between the corporate LAN and the remote location of those end users. Connectivity still uses the internet for access, but the primary difference is that a hardware appliance.
is used on both sides of the VPN tunnel for automated authentication and encryption across the virtual tunnel. The benefit to the end user is that they are not required to manually authenticate each time they need to access a company resource. Instead, a site-to-site VPN acts as if it's simply an extension of the corporate LAN.
Previously, the high cost to deploy and remotely manage dozens or hundreds of site-to-site VPN tunnels led many IT departments to use site-to-site VPN deployments sparingly. But thanks to lower hardware costs -- and advancements in cloud management technologies -- offering static VPN tunnels to large numbers of teleworkers is now a reality. Several examples of this exist in the market, including the Cisco Meraki Z1teleworker gateway appliance that offers a low price point and a cloud-managed interface for ease of troubleshooting by corporate IT staff, as well as entry-level appliances from Fortinet and Checkpoint.
Finally, if you need traditional remote access services but would rather have someone else manage the entire architecture, you can go with a fully managed VPN provider. In this scenario, cloud-based remote access is achieved by allowing a cloud service provider to not only manage authentication but also the authorization, accounting and general maintenance of a standard remote access VPN service. Plenty of service providers offer VPN as a service including technology companies like MegaPath and Zscaler. Wireless carriers such as AT&T and Verizon also offer business-class remote VPN access services that primarily target mobile workforces that use smartphones and tablets to reach corporate resources.
by Andrew Froehlich
Questions? contact us