Demand for predictive intelligence technology is on the rise, in part because of the incredible results that businesses are seeing when they target consumers with unique messages based on searches, clicks, views, shares, comments, and downloads. According to a Salesforce report, predictive intelligence yields a 40 percent influence in revenue after 36 months of implementation, and the average lift in conversion rate is 22 percent.
Small business owners have the tendency to shy away from advanced technologies like predictive intelligence, however experts in the field say that’s a mistake, and many of today’s platforms can be implemented by merchants on Main Street.
“Because most SMBs rely on small marketing teams, predictive intelligence is actually a strong strategic choice for them,” says Gordon Evans, vice president of product marketing at Salesforce Marketing Cloud. “Advances in technology have enabled businesses to implement predictive without the need for that team of data scientists — putting data science directly in the hands of marketers and letting small teams have an outsize impact.”
Here are seven ways that small businesses can get in on the action and start using predictive intelligence tools today.
1. Start with a realistic goal. “Businesses need to first pick a predictive use case. In other words, what are they trying to predict? Is it sales for a given product on inventory next month? Is it the likelihood for a repeat customer to churn? Is it what products go with a given product? Their creativity is the limit. Once they decide on that, they can export data out of their POS systems and onto a spreadsheet. In the spreadsheet, they can reorganize columns and create new calculated columns — such as average spend per customer — as long as they think these additional variables may correlate with their target variable. This is where domain expertise is key.” (Atakan Cetinsoy, BigML)
2. Offer increased personalization. “Small business marketers can take advantage of predictive intelligence by utilizing it to offer up increased personalization — for example, utilizing predictive audience segmentation to automatically divide consumers into targeted audience segments, based on their predictive scores, that funnel them into specific outreach campaigns. Customers who might be likely to disengage with the brand will be marked to receive win-back campaign outreach, such as emails with discount offers.” (Gordon Evans, Salesforce Marketing Cloud)
3. Use insights as input into event planning. “Customers are able to use the data from a predictive platform to create and execute very targeted ABM [account-based marketing] campaigns. Some customers have been able to better target their events. They used predictive intelligence to develop a scored list of target accounts. Then they overlaid those accounts and scores on a Google map to see where the high-scoring accounts were clustered. They used this insight as an input for their event planning, so the decision around ‘where should I hold my event?’ becomes a lot more data-driven.” (Nipul Chokshi, Lattice Engines)
4. Bid strategically when advertising online. “Small businesses can take advantage of machine learning by advertising on platforms like Facebook Ads or Google AdWords and using the automated bidding feature. This will determine exactly how much a marketer should bid on a given ad space, taking into account variables such as demand from other marketers, the desired demographic, and the value of the particular user being bid on. These platforms can predict which users will be most valuable to a particular marketer based on desired outcomes, such as online sales or website visits. Over time, these predictive capabilities improve thanks to the platform’s machine learning capacity.” (Anthony Long, Vistaprint)
5. Evaluate each customer’s likelihood to convert. “Predictive intelligence can help small businesses gain the competitive edge they need by making the most valuable use of their limited time and resources. For example, take predictive scoring — the ability to use data science to learn and score a customer’s likelihood to engage. By identifying which customers are most likely to take actions — such as opening an email, unsubscribing from an email list, or making a purchase — SMB marketers can focus their efforts on the customers who are most likely to convert.” (Gordon Evans, Salesforce Marketing Cloud)
6. Link together customer interactions. “You have to start with data that links together various interactions that you have with customers — think loyalty programs or opt-in email notices. Insightful analytics on that data don’t have to be hard. Tried-and-true recency, frequency, and monetary (RFM) analysis works great to find the highest-value customers. Combining them in a weighted score works well. Then use that to understand on whom you should spend more time and resources. Analytics doesn’t have to be complex to have value.” (Ken Inman, Neustar)
7. Close the loop on your model. “Do close the loop on your model by putting it to the test against new data it has not yet seen. Some models are great with historical data, but they fail to transfer this success to new instances due to something called overfitting. Predictive intelligence tools let you split your historical data into test and training subsets, so you can ensure you have a validated model before you put it to use. Closing the loop also means measuring the business impact of your predictions on a regular basis, be it ROI, cost reduction, or increases in sales.” (Atakan Cetinsoy, BigML)
Stephanie Miles is a senior editor at Street Fight.
Interviews have been edited for length and clarity.