by David Rogers
That might seem like a strange way to open an article on artificial intelligence (AI) and marketing. After all, can’t we just ask a computer to tell us what makes marketing effective? But the truth, as we’ll see in a moment, is that computers are only as smart as the data they’re given to learn.
Which is why this is such a critical question to ask, and why I want to answer it up front. Effective marketing – the kind that leads to sustainable growth for a repair shop – is marketing that attracts and retains quality customers. It builds trusting relationships with loyal customers who maintain their vehicles, refer friends and family, and come back year after year.
Effective marketing isn’t measured by ROI or car count (no matter what some marketing companies tell you) or by how busy you feel (no matter how much stress it takes away), but by looking at the long-term health of the business and loyalty of your customers.
And when it comes to AI, that difference in the definition of effective marketing is like a chasm. Here’s why.
First, let’s remove some of the mystery. AI isn’t sorcery. You can’t just turn on a computer and wait for it to become self-aware. You must teach it, and it learns a little, and then you test it and teach it some more.
Think about self-driving cars and a simple stop sign. When a human is behind the wheel, they can recognize a stop sign easily. It’s got eight sides, it’s red, it says stop. If there’s graffiti or a sticker on the sign, it’s still easy to identify it as a stop sign. But if you’re a computer, and you expect a sign to be red and octagonal, say stop in white and now there’s other writing on the sign, does it still mean stop?
What if it’s dark and it’s hard to tell if it’s red? What if it’s partially behind a bush? Is half a stop sign still a stop sign?
AI is the long, drawn out process of showing a computer enough pictures of stop signs – big, small, clean, dirty, easy-to-see, partially hidden, and on and on– that the computer can recognize that simple stop sign and obey basic traffic laws.
But what happens if you teach your self-driving car that stop signs aren’t red and octagonal, but yellow, upside-down triangles? Just because the computer thinks that yield signs are stop signs doesn’t make it real. It will now make decisions based on bad data.
Which is why it’s so critical to be clear on the definitions of marketing and AI. If you teach a computer that effective marketing is based on car count, then you’re likely to get results that lead to more car count at the expense of the size of your average ticket and customer loyalty and long-term success.
In other words, AI isn’t a magic elixir. Indeed, there’s plenty to be wary of.
Not least of all because “Artificial Intelligence” is a marketing buzzword right now, and can be twisted to mean something as simple and basic as “we use spreadsheets and graph the data.” Real machine learning — the kind that leads to better marketing — cannot happen overnight. You must spend the time measuring marketing so you can teach the computer what effective marketing is and how it works.
Even more importantly, and to hammer this point home, it also takes an understanding of what makes marketing effective.
In other words, beware of anybody who promises AI-driven marketing that doesn’t have the history (and therefore the data) to make that claim possible. There’s a reason why people get degrees in statistics and mathematics to start a career in teaching a computer how to learn – it all starts with a lot of measurement.
Better marketing messages
We are not yet to a point where computers can start crafting marketing messages. There’s a reason why the internet is full of hilariously terrible attempts by AI to write marketing messages, movie scripts and songs: the technology is not there yet.
By measuring the effect of a marketing piece on the health, trust and loyalty of your customer base, a computer can recognize what marketing messages work best with particular types of customers. Imagine being able to target great customers with the messages that are most likely to work on them, every time.
But that big warning light is flashing again. You can train a computer to tell you which marketing is most effective, but if you don’t know yourself, the results the computer provides will be ineffective. If you train a computer that customer loyalty is less important than an immediate spike in car count, AI will give you the tools to destroy your customer base more effectively than ever.
Better marketing budget
We all know auto repair is seasonal. You can mitigate it through your marketing, but summer is always busier than the holiday season.
What we’re finding is AI can help plan for that seasonality.
There aren’t shortcuts, of course. It takes looking at years of past performance, and years of marketing budgets, and again, understanding how to attract and retain quality customers, but AI can use all that measurement to help plan for the future.
Better marketing results
When you target the right customer with the right marketing at the right time, you can sustainably grow your shop. But, just as AI doesn’t happen in a vacuum – it must be taught the real results of effective marketing to predict how to increase those results – your marketing doesn’t happen in a vacuum either.
You can send the perfect message to the perfect potential customer, but if your customer service breaks down at the shop, you’ll lose that customer forever. Unless you’re also measuring your team and your shop, unless you are fixing production bottlenecks and perfecting your service advising processes, unless you’ve got your profit margins dialed in, your incentive pay plans in place, and your team trained and accountable, marketing can and will fail.
There’s a common trap that shop owners fall into when it comes to marketing: “if only I had more cars, all of my problems would be solved.” It leads to owners switching marketing companies once a month to try something new and/or less expensive. That means havoc for marketing, of course. If no campaign can ever be maximized because it doesn’t run long enough to be measured and adjusted, then a significant portion of those marketing dollars end up wasted.
But more than that, it ignores the problems that led to low car count in the first place. Yes, sometimes, more cars are all that’s needed. But more often than not, car count dropped off for a reason. Poor customer service, incomplete inspections, broken production pipeline, uncommitted team…these things must be measured (even analyzed by AI) and improved if your marketing is going to be truly effective and sustainable.
In other words, AI can lead to a marketing renaissance for auto repair shops, but it can’t replace doing the important things that lead to a shop being successful: measuring your team, training them and holding them accountable, every day.
This is an exciting time to be a shop owner. AI is starting to give us control over every aspect of our shops – from accountability to training to marketing. If you understand who is behind the AI – and how they’re training it – the coming months and years will change the industry for the better!