Influencer Marketing: From Mystery to Data Science
Traditional advertising has faced a sharp decline. People are looking for recommendations from other people just like them. They don’t trust an actor dressed as a dentist anymore; they want to see a real person walking into a dentist’s office and coming out with a bright, white smile. Thus, the era of influencer marketers is dawning. The definition of an influencer is somehow blurry, since, depending on the niche, it can be anyone from a teenager talking about make-up to high-profile tech evangelists.
Although brands are beginning to notice the advantages of using this kind of marketing, putting precise numbers on the ROI charts is still a challenge.
Challenges of budgeting influencer marketing
The finance departments of companies are very reluctant when it comes to spending money on influencer marketing. Even if this is a new marketing tool, it still needs to fit into traditional marketing budgeting plans.
Their worries are not unjustified since until recently, it was difficult to link the money spent on this type of advertising to the actual sales.
The early indicators included tracking the number of impressions posts got. However, this did not prove that the viewers were interested in the brand, or even if this interest was positive or able to generate sales. A mild improvement was tracking the actions of the followers, including likes, shares, comments or any other way of showing that you care about the content. This meant only a minimal chance of getting to the end of the sales funnel. You can like a car on social media, that doesn’t mean that you will get one any time soon.
How can data science help?
To answer the money riddle described above, marketers have turned to data science for more precise answers given by statistical estimations based on analyzing large volumes of interactions.
Measuring the audience
Just looking at the number of followers a profile has is not enough to classify them as a good influencer for your brand, even if the numbers are impressive. You need to know other data like the in-depth segmentation of that audience. Are they from a location you are interested in? Are they the right age for your product? Don’t just assume that the area of the influencer overlaps with their target audience.
Analyzing the content
Not only is their profile of followers important, but the quality, tone, and originality of the material are factors worth considering when choosing an influencer.
Working with a platform which is powered by data science can analyze the performance of each post in real-time. This tells the marketing team what kind of content gets the right response from the target audience and it’s a good starting point in developing a strategy for later.
Find the right influencer
The first question when thinking about an influencer marketing strategy is finding the most relevant people to work with. This should be a combination of different demographics, interests, reach, the audience type and size. Putting together the information from the previous point should act as the base for pinpointing the right person for the job.
How does that work?
Automated systems can scrape the web based on a few keywords or hashtags and identify the profiles which rank for those indicators. After this step is complete, the next one if to look for the most suitable influencer. This is done based on the demographics of the target audience.
Direct measurement strategies
A recent study on influencer marketing discussed by Forbes highlights the influencer’s impact on different classic marketing efficiency indicators including basket size, sales lift and even age-old techniques like promotional codes redeeming, in a new context.
The most accurate way to identify the impact of a campaign using an influencer is promotional coupon redemption. This technique has the advantage of penetrating the entire sales funnel, from acquisition to purchase. Furthermore, even if this technique is older than a century, it can still be used in combination with social media influencers. It can establish a clear baseline for the performance of each influencer and serve as a tool in the decision to reuse a specific influencer. Also, this approach will satisfy the finance department entirely, since it can put a real number in the ROI column.
Loyalty card data measurement
An imperfect way to observe the efficiency of a specific influencer is to analyze the variation of purchasing through loyalty cards. This method should only be applied to the people in the target group, taking as a control group those who don’t correspond with the audience of the influencer. Such a comparison can determine if the increase in sales came from the exposure to the recommendation, or another factor such as seasonality.
Probably the least relevant way to measure the impact of an influencer’s campaign is to compare POS sales inside and outside the campaign. Even if there is a particular variation, this is hard to be attributed to the influencer unless there are no other undergoing promotions at that moment.
Talking about brick and mortar locations, a simple survey could show if the influencer campaign had any impact on the purchasing decision. Online, the same information can be easily retrieved by attaching a pixel to the post or using specific tools.
The key takeaway from this analysis is that although difficult, influencer marketing can and should be measured. As the tools evolve this will become the standard practice and self-titled influencers will be eliminated.
Also, signing a high-profile influencer to do one post will most likely not be enough and not yield the expected results. As in the case of traditional marketing, people need repeated exposure to the same content to go from awareness to purchasing. In this case, using more medium-sized influencers could prove to be a better strategy.