Your B2B company invested millions into a new Data Science or Analytics COE. To prove their value, the teams outlined ten data science workstreams in Year 1, ranging from Route Optimization, Staffing Optimization, Price Optimization, Inventory Efficiency Improvement, Actioned Analytics for Customers, and others. The projects spanned all critical functional areas and collectively had a 30x ROI associated with them (in other worse, the projects would deliver 30x more Operating Profit on an annualized basis compared to the cost to maintain your Analytics COE).

One year has passed, and you find that the Analytics initiatives have delivered on most objective outputs. All the data tools were deployed for Pricing, Supply Chain, Finance, and Marketing teams, and most product management metrics were met or eclipsed (# of data pipelines built, # of models deployed, # of feature requests completed, etc.).

There's one problem: adoption of your analytics solutions has struggled to take off, and few measurable outcomes, if any, have been achieved.

As my good friend Pratik Kodial, former AI/ML executive and currently a successful hedge fund manager, pointed out:

"[the frequent failure of analytics initiatives] doesn't fit the narrative sometimes leaders want to tell the markets, and what the markets want to hear."

The above has been endemic in the field of analytics. While it has gotten slightly better over the last five years, it still has ways to go before there's a direct, measurable linkage between analytics initiatives and incremental financial performance.

There are many causes, one of which is that nascent Analytics / Data Science COEs often overcommit, usually in response to overexcited CXOs who think that AI/ML can be the panacea to many business ailments. 

It's a delicate balancing act for Analytics leaders: 

1. How do you manage the internal or market perception and entertain bold (but often unrealistic) CXO ideas around Digital Transformation versus

2. Leading your Data Science team to successfully deploy smaller, more focused initiatives that will generate measurable returns for the organization.

The former ensures that your team gets executive recognition and funding for growth (often the best strategy for self-preservation), while the latter sacrifices bold thinking and massive initiatives in favor of more pragmatic, value-added solutions that decision-makers will use and drive value with.

If you are a new Analytics Leader, your job is to convince your CEO that doing #2 is the right strategic path for the first two years.

If I were to counsel new Analytics / Data Science COEs (or new leaders stepping into old teams), my advice would be the following for the first two years:

1. Focus on fewer initiatives that directly impact the Revenue and Gross Profit drivers of the business.

2. Prioritize a subset with the right balance of Impact, Level of Effort, and Organizational Support (at the Manager and Director, not just from Sr. Executives).

3. Figure out which projects can be outsourced to expert firms who have done this successfully for a while vs. which projects should be done organically by your team (the buy vs. build dilemma).

4. For each analytics project, leverage an internal Core team of experts from relevant functions and various levels (Analysts, Managers, Directors). Build your analytics solution in a way that involves the Core team at each key stake - ultimately, you will build something that is just as much "theirs" as yours. 

5. Don't write your first piece of code until you understand the problem and the value of solving it and have split the problem into smaller, manageable components that you can attack.

Doing #1-5 in B2B usually means that you'll end up with a list of 4-5 key Analytics initiatives around the following tactical areas. 

1. Price Optimization: Helping your Sales and Revenue Management teams set better prices that grow Net Revenues while growing (or not sacrificing) Gross Profits.

2. Customer Churn Reduction: Ensuring your Sales team can be proactive and surgical about customer outreach to minimize Churn and increase overall revenue.

3. Cross-Sell Optimization: Giving your Sales teams facts- and often ML-driven, actionable insights on cross-sell opportunities for high potential Customers based on historical transactional panners and cohort analyses. 

4. Promotion and Discount Optimization: Helping your Sales teams make more bonuses by increasing their Net Revenue and Gross Profit performance through smarter price promotion and discount management.

5. Product Procurement Optimization: Helping your Merchandising team buy the right products in the right Quantities. It is often the #1 cause for Distributors and Retailers to end up with lots of unproductive inventory in their warehouses. 

If done correctly, the above analytics initiatives will keep your team busy for the first couple of years and ensure internal adoption and value generation for subsequent years.

Once you've maximized the incremental value from these Sales & Operational Efficiency type analytics initiatives, it's time to focus on the large-scale, Digital Transformation type efforts. By that time, you will have proven that your team is vital to the Company's financial performance and will have greater comfort in managing through the complexities of mass-scale initiatives that carry substantially greater financial and reputational risk.