Digital B2B marketing obsesses over first-party data.

Everyone wants more leads, more firmographic and psychographic data sets; it’s a maximal approach to data acquisition born from the harsh reality that a marketing department’s influence lies primarily within upper-funnel strategies.

But for marketing organizations that cultivate more collaborative relationships with Sales Ops functions, no other first-party data set holds higher priority for integration than sales data.

According to McKinsey, data-driven cross-selling can lift sales and profits by 20% and 30%, respectively. Whether you’re segmenting your 1:1 marketing approach based on annual account values or transactional level data, here are some principles to keep in mind for your organization to capitalize on cross-selling.

Right place is easy. Right time is trickier.

While there are plentiful applications for audience building on social and programmatic media platforms, first-party sales data is far more likely to reshape a marketing team’s approach to the 1:1 channels with minimal upfront activation costs: email, SMS and in-app notifications.

At this juncture, knowing where to meet your shoppers isn’t too tricky, but knowing when to engage them requires more data analysis.

For example, we recently developed an always-on lapsed buyer campaign for one of our clients. The campaign’s core communication was a time-to-restock email that would be deployed to a customer who hasn’t purchased product for a certain period.

The challenge was the “certain period” is not a one-size-fits-all window; different customers consume products at different rates. They also purchase products at different volumes per transaction. To ensure the campaign’s effectiveness, our data services team analyzed hundreds of thousands of transactions to calculate a “lapsed threshold” period personalized to each specific customer.

This approach ensured we would reach each customer with a time-to-restock message on a timeline based on personal behavior, not a broad aggregate. The results of this approach were fantastic. Users who engaged with our communications spent an average of +$287 per transaction compared with those who didn’t.

Know and control your external variables.

Similarly, it’s vital to recognize important external variables when launching work on campaigns driven by first-party sales data. At the simplest level, campaign measurement is most effective when you withhold a statistically significant control group from the marketing tactics. This ensures more accurate assessment of the 1:1 campaign’s influence and impact.

This also requires awareness of marketing tactics that may cross-influence your 1:1 campaigns. Retail- or program-specific discounts may subtly influence buyer behavior outside of your control group. Depending on the specificity of analysis, seasonality may play into buyer behavior as well. That’s why we recommend investing a full calendar year into analyzing buyer trends to ensure time-relative criteria can be shrunk or lengthened to accommodate peak seasons and slowdowns.

Don’t let AI lead your data analysis.

Given all the current hype, trying to find objective research on machine learning’s efficacy across various B2B operations is tough. This is partially because the AI industry is still nascent but also because the technology’s efficacy varies wildly from function to function and platform to platform. Today, too many marketers leverage AI like it is a clutter-free Google – a search engine alternative to assist with code troubleshooting or A/B testing ideation around an experience. Yes, AI holds incredible promise and can be truly useful, but only if you really know the platforms along with their strengths and weaknesses.

Conveniently, it returns just one simple answer to every user’s question.

The issue is that indiscriminately-sourced AI answers are wrong literally most of the time. According to a 2023 Purdue University study, ChatGPT returned incorrect answers to approximately 52% of questions its researchers posed. By any measure, that’s unacceptable.

Simply put, AI-generated data insights should not be interpreted without a deep familiarity of the data sources being digested and the platforms being used. Your sample size may be too small to produce valid conclusions, but AI won’t tell you that; you need to apply that judgment. Are there inconsistencies across your data files? AI won’t warn you about those either.

Before you turn to AI as a silver bullet solution to draw hypotheses from your data, make sure you are using the right AI source.

So even if you believe you’ve exhausted your analyst team’s ability to draw hypotheses from your sales data, do not look to AI as a silver bullet solution. It will only make your analysis (and future AI interactions) suffer even more.

Better still, ask the humans at Bader Rutter. About analyzing data or intelligently engaging AI, we can definitely help.