One of the main takeaways from Salesforce’s 2017 STATE of MARKETING report is that investments in AI has outpaced spending in other marketing tech areas. B2B marketers are adopting AI technologies ranging from predictive lead scoring to chatbots in droves. But before you get caught up in the hype, there’s one thing you need to nail down before you start applying AI into your marketing processes: Is your B2B contact leads database ready for AI at all?
To answer this, we first need to separate the reality and the publicity behind AI’s capabilities in B2B marketing today. MarTech Advisor points to four key areas where B2B marketers can realistically expect AI to lend them a helping hand:
- Scoring and ranking leads.
- Segmentation and content personalization
- Discovering and implementing Marketing automation strategies
- Sales enablement and acceleration
At its present development stage, the best that AI technology can do is allow you to carry out the tasks in each of the above activities more efficiently. While some aspects of AI can uncover prospect behavior invisible to the unaided human B2B marketer, the reality is that AI remains just a tool, and tools are only as effective as the persons and processes using them.
So if you think AI has a place in your marketing toolkit, you first need to take a good look at your B2B contact leads database.
Like everything else in marketing, AI depends on good data. The data currently sitting in your CRM and datasets you’re about to collect need to meet some basic requirements before starting AI-enabled campaigns. In an interesting video series, Brandon Rohrer at Microsoft Azzure thinks of data science and AI as a lot like making pizza: the better the ingredients (your data), the better the final product (marketing insights).
There are four qualities that any dataset must satisfy to be ready for AI and data science:
- Relevant: Do the fields and records in your B2B contact leads database help you answer the questions you’re exploring? For example, which lead attributes in your CRM influence the likelihood that a prospect turns into a customer within the next quarter?
- Accurate: How reliable are the models/profiles generated from your marketing database? Do the records contain incorrect, outdated, redundant, or invalid entries?
- Connected: Are there significant gaps in your marketing data? What percentage of records contain empty fields?
- Sufficient: Do you have enough records to build robust AI models?
While each of the above criteria is important, we need to carefully consider sufficiency. AI requires data–lots of data. The algorithms that power most AI applications run on vast amounts of examples in their training set. In general, the more examples you use to train an AI algorithm, the more accurate the resulting model gets.
So before you think about applying AI in marketing, you first have to bring your contact leads database up to snuff. Use the previous ideas as your guidelines and maximize the power of artificial intelligence.