Artificial intelligence (AI) is everywhere, and marketers have been busy spreading it on thick, like barbeque sauce on ribs. But why are ribs delicious? They use the right sauce. Layering on the wrong flavor or using the improper amount of anything is always a bad choice. AI is no different.
Just because it's the topic of the time doesn't mean it's one-size-fits-all, and it also doesn't mean a formula that works for someone else will work for you and your needs. To recognize a strong AI return on investment (ROI), you need the right recipe, which comes down to a tailored strategy.
Marketing benchmarks
There are many benchmarking resources marketers can use as a starting point to calculate potential ROI for marketing campaigns. For example, you can reference Mailchimp if you want email open rates by industry. You can also check out Wordstream's resources for Google Ads and Meta paid social media to set your ROI goals. You can even look at a more comprehensive view of marketing metrics like sessions, average session duration, engagement rate, session conversion rate, impressions, position, clicks, click-through rate (CTR), searches, views, interactions, and website clicks from a vendor like Databox to get started from solid ground.
At Imarc, we use all of these resources, as well as competitor data, third-party research, and historical performance data from our clients, to set goals and, ultimately, decide if our efforts have paid off.
AI ROI in marketing
We haven't personally used their products, but burgeoning education AI chatbot content shows a wide range of ROI possibilities for companies. Verge AI was built for student engagement and they estimate that an AI chatbot can generate up to 4,158% for universities. However, that's specific to the education sector, depends on how it's implemented, and is based on a hypothetical scenario.
And despite messaging in the marketplace saying it's easy to implement AI chatbots, demos don't necessarily translate to the real world. Custom enterprise-level chatbots are challenging to build as they reference internal documents to send answers to users.
When it comes to using AI in marketing, the benchmarks don't exist yet, at least not in any meaningful way. You must dig deeper to decide if using that technology can work for your strategy and generate the right level of ROI you want.
Here's a good place to start.
Define why you want to use AI
We use AI for our clients and our agency when it's the best solution for a problem. A premier example would be Skillsoft.
Recently, they asked us to help them devise a tool to save time with language translations. Given the minimal investment and potential for maximal payoff, both parties saw great potential for an AI-driven solution for this type of tedious work.
The choice to integrate ChatGPT into Skillsoft's Craft instance took them from three hours to just three minutes of production time per landing page and, monetarily, from a $120 cost per page to the low amount of $1.85.
You can read more about how we saved Skillsoft money and time in our case study.
But this isn't about being promotional. This is about finding the right problem to solve using AI, not just using it to use it.
Create a strategy & execute
Once you find that opportunity, backtrack from the result you hope to achieve and outline determined steps to take you there. Along with these steps, carefully inventory costs that pop up to do with your AI project. These costs can be used later to see if your plan worked!
At the same time, now is the moment to note what an outcome, whatever that may be, is worth to you. If you're a recruiter, you know what your benchmarked cost per hire should be and what it would mean if you could beat that cost. Within your respective industry, figure that out and put it into your plan.
Finally, you need to consider all the potential obstacles. How will your AI tool be governed? What compliance rules must you follow? How will you protect your data? Do you have clean enough data to leverage AI effectively? Are you in this for the long haul (AI implementation can take longer than you might think)? Who will own the tool, and who will help support it? Answer all the questions you can upfront so you don't get surprised later on.
When you've done all these things, create the tool and launch it!
Be patient
Trust the process. If you do choose to add AI into your strategy, it's not necessarily going to be a fast fix. Sometimes, web engineers have to fight with the AI. For example, it's possible to ask AI to generate code that completes a specific task, but sometimes it doesn't work. Instead, it produces an error message, and it can mean getting stuck in a debugging cycle. Occasionally, the best solution is to start over and use another prompt. In addition, off-base prompting can overcomplicate a solution, so you have to re-prompt to simplify the codebase. Context is everything with AI (as it is with most marketing). The more context you give the large language model (LLM), the better it can help.
The point is that trial and error takes time, and you're going to want to do this right.
Measure
To calculate your ROI, most use this basic formula: Marketing ROI = (Marketing Value − Marketing Cost) / Marketing Cost. Take all the costs you incurred during your strategy and execution phase to determine your marketing cost, and then stack them up against the value you recognized. Even if your AI tool isn't a great investment right away, if it still has potential over time, you can re-measure after you improve your strategy or update whatever you've built.
Optimize
Once you've determined your initial ROI from AI, take your learnings and decide whether to set your innovation aside or dig in deeper. In paid media, your first attempt is rarely your last, and with AI, some experimentation may be necessary to see your ultimate reward. Only you can decide that, though.
AI, the right choice for your company or a trend you're hoping to jump on?
According to Forbes, "a survey by McKinsey [found] 77% of companies are either using AI or exploring its potential." Still, the same data shows that "fewer than 20% of organizations achieve significant financial returns from their AI initiatives."
That's a whole lot of people investing a whole lot of money for – not a lot of ROI. Why? Because they aren't supporting their case for AI with deep thought from humans. AI will never replace that kind of thinking, and it isn't meant to. AI works best as a boost to the power of the human mind. It can send us further, faster. It allows us to spend more time solutionizing because we have support for menial tasks.
We're not saying you shouldn't implement AI. We're saying you should use it purposefully to impact your specific business. This kind of strategy will generate the ROI you want.
Unilateral AI ROI is a myth, but deploying AI strategically is the present and future. If you have a challenge you think could be solved with AI and you'd like a partner to help, say hello.
