The idea of automation is to reduce the amount of time you or employees spend on tasks. This saves money, but well-executed automation can also lead to more reliability, accuracy and greater potential.
It’s thought that Facebook business model is essentially based on PPC ads and the underappreciated chatbot for each company’s page. Chatbots, whether it’s an AI, self-produced one or Facebook’s, can help improve CRM (Customer Relationship Management). You can essentially feed the customer towards the correct sales funnel, or towards the correct resource to help them solve a problem.
Not only does this save a bunch of time, seeing as we’ve tricked a rock into doing it for us, but it’s actually more helpful to the customer in many ways. Most importantly, the response time is instant, compared to a human answering which has to read and type out the answer – and that’s if there’s no queue. This leads to better customer engagement; they’re sure that a quick message to the company for some help will be awarded instant attention. It’s also easier to approach global markets this way – chatbots can be programmed into different languages, something that isn’t viable with a customer relationship team.
Billboards are still being used around the world to advertise to the public. What’s more common, though, is billboards/posters on your own premises. For example, the shop window outwardly facing a huge “35% off”. LED screens are such a worthwhile investment, because it reduces all of the future printing and setting up costs of new images/designs. This makes it affordable – but is far from the core reasons of getting one.
Outdoor LED screens are simply far more effective. They’re not just brighter and more eye-catching, but they’re customizable and automated. You can set up 10, 20 or 100s of designs ready for display. Along with this, you could pair each with its respective target audience. For example, “25% off men’s suits” and “free toy with every purchase” are for two different messages for a big multi-purpose store. The first could be scheduled to appear on the screen from Monday to Friday between the times 5PM and 7PM, because this is when commuters pass the store and may need a new suit. Likewise, you could target the free toy message on only certain screens – perhaps the one that is situated on the east side of the store because that’s nearest to the Disney shop is.
One thing that LED screens can’t manage is a customised advert based on who walks past in that moment. This is what online adverts are based on, of course, and how automation can be leveraged to really personalise adverts.
By creating (or gaining access to a 3rd party’s) customer profiles, we can begin to build up an array of knowledge which relates to the customer. Simple things, such as their name, as well as more detailed things, such as their living situation, their hobbies, or even their behaviour patterns.
The more you know about someone, the more bespoke the message can be to them. A hypothetical example could be a Tesla car which stores information on your driving patterns. Knowing you like to stop for fast food on Fridays, because your car’s GPS goes there, and because Tesla have a hypothetical partnership with McDonalds, you may suddenly be getting McDonald’s adverts when listening to the radio in your Tesla. Even worse, you may find it becoming a factor in how Tesla designs the routes that it will autonomously drive past.
The reason for this extreme hypothetical scenario is to point out not just the effectiveness of personalisation through algorithms, but also the ethical concerns.
Social media posting
As a marketer, writing 4 posts per day for the company Twitter and Instagram account, spread exactly 2 hours apart from each other, can really get in the way of other work. It’s much more efficient to simply write those four posts in one go, then schedule the posting of each. This is just one way that automation helps with social media, but the reality is there’s no end to the possibilities.
One unique application of algorithms to social media is sentiment analysis. Here, you can find yourself quantifying (through a pre-built model on, say, Python, and facilitated by an API) the emotion of social media posts. By assigning numbers to sentiment words (i.e. -2 for dislike, -4 for hate and +5 for love), or by using ML to produce the lexicon, you can find the aggregate reaction to a topic. This could be great for A/B testing products — simply create a post about the product, but choose two different colours for each post, and see which gains a more positive