Client’s Problem: “We really just want to know what the future holds for sales. That knowledge impacts everything we do.”

This problem is so common that there are thousands of off-the-shelf solutions. All the big companies (Salesforce, Tableau, SAP, MicroStrategy, et al.) have robust forecasting tools baked into their desktop or cloud software that only require the press of a button. During scoping, we pressed the Client hard on the validity of their “build versus buy” decision. We offered our insights and assessment as well. Still, they had several excellent strategic reasons for wanting to create their solution.

As with any problem, we searched the Client’s data and their industry data for patterns. We found no shortage of those: temporal, spatial, segment, category, and everything. What we found were too many patterns. Something didn’t add up: we were looking at a single data set despite the data sets being independent. The multitude of patterns was the signature of a coupled system with a driver. So, we unearthed the small number of driving forces in the system.

Once we knew the underlying drivers, we modeled them using traditional statistical techniques and modern machine-learning models. The models accurately reflected the Client’s desire for a time horizon relevant to their decisions. Mission accomplished! However, was it?

Using the knowledge we gained about the demand patterns in their industry, we saw a unique opportunity for the Client to acquire, collect, and curate novel data associated with their industry. As the Client’s liaison, we worked with third parties to design several strategic initiatives to tackle data acquisition. The new data was later folded into the demand modeling as market share features and further improved the model’s utility as a strategic tool.

Client’s Problem: “Our sales transactions are pretty rich with data, and we don’t have any trouble reporting on that data or using canned solutions from our CRM provider to determine customer lifetime value (CLV) or predict churn. Where we’re struggling is in determining how to engage with a customer next in order to maximize our market share, not to maximize their spend necessarily, just their number of purchases. A lot of the products we sell you can get anywhere, and we want our service model to shine by getting people regularly into our sales funnel.”

This was a “fascinating” problem, which concerned us, it didn’t seem very valuable to the Client at first glance to have a consultancy dig on this problem. Market share is difficult to ascertain if the Client is not in an industry with numerous public companies that disclose data where one can reverse-engineer the market share per company. That was the case here. However, what ultimately made a solution tractable and valuable was that purchases were geospatially related.

A common approach to customer behavior modeling is to consider the purchase network or behavior tree. Customers in the Client’s industry had a set of physical limitations that constrained their purchases, adding to the behavior tree a unique set of conditions directly related to market share. As a result, we were able to build a probabilistic model for how many purchases a customer should make in a period. Combined with the Client’s existing churn and CLV models, the behavior model was able to discern missing purchases from non-existent purchases.

After segmenting and clustering the Clients’ customers, we probed clusters for actions that resulted in long-term growth in the number of transactions. We subsequently built an intelligent agent using reinforcement learning to execute the next best actions following a purchase with a high probability of increasing future transactions. The Client’s sales team used this agent to reduce the amount of unnecessary “discounting” (e.g., coupons, bundles, sales) to attract continued purchasing.

Client’s Problem: “We operate in a hypercompetitive industry where our products have slim margins, and the differentiation within a market is service-level and not product-level. Our issue is not knowing how competitor behaviors and strategies impact our sales. We struggle with the basics, like pricing, to capture the value proposition of our service over competitors. Even simpler things, like when and how to execute product decisions, are poorly understood outside of the tribal knowledge of our product managers, making long-term strategic decisions extremely difficult. All of this work is also very labor intensive, and the drudgery of it all distracts our talented people from doing other, more valuable work.”

We have a couple of former product managers on our leadership team, a widespread problem for retailers of any scale. Most companies avoid the problem of competitor intelligence altogether by washing it out (sometimes rightly) using one of the standard product management policies, e.g., margin or premium. During scoping, we determined there was sufficient evidence and data to pursue a more sophisticated approach, which was ultimately adopted.

The Client was a large national company, and they did indeed have many competitors. Our first step was to use the Client’s existing industry data to determine if there were patterns in sales, demand, and competitor data. We found signatures of competitor behavior impacting sales that were well-described by game theory. With this discovery, we automated the codification of the rules and actions seen in the games to build a data set for modeling. Using this game characterization data, we created machine learning-based models ensembled to predict when, where, and how a competitor would make moves to capture demand.

Interestingly, these models performed poorly during the experimentation phase. We learned in the experiments that an active spatial component to the competitor behavior was not evident in the original data. Competitors also had complex but quantifiable interactions over time, indicating structured network effects. We modeled both patterns in tandem using a novel approach based on neural networks. We automated the entire modeling backend into an intelligent agent for interaction with our Client’s analysts. The speed and efficiency of this new capability enabled rapid testing of sophisticated strategic scenarios they had never contemplated previously.

Client’s Problem: “Our analysis department is overrun with work, and our business partners don’t have the people or resources to self-service. Our people end-up being managed through their email inbox with no coordination, meaning frustrated business partners, duplicated work, cut corners (tech debt), no time for innovation, and low morale. How do we fix this?”

This client had a process problem. During the scoping and discovery phase, we interviewed the analysts – all of whom had plenty of enthusiasm, talent, and skill. Likewise, their business partners possessed the needed subject matter expertise and strategic vision to facilitate high-quality analytics work.

The disconnect occurred in work coordination and precisely how the analysts worked together. There was no “strength in numbers” credo or prevailing spirit of “we succeed as a team, fail as individuals.” To rectify this issue, we designed a workload management process that blended agile methodologies and traditional governance.

As the first step, the analysts self-selected groups of cross-functional “squads” with business partners that had overlapping strategic needs. We also created an informal “capability owner” role to help each squad take strategic ownership of its work. These squads were taught a blend of agile techniques (backlogs, sprints, stand-ups, reviews, retrospectives) and encouraged to modify them to suit their needs.

Additionally, the department and executive teams designed and implemented a process for managing and governing work requests to separate the distinct forms of work we found: PMO-governed strategic projects, support, and ad hoc analytics. The outcome was high-performing analytical teams embedded within the business. These teams leveraged agile methods to manage their workloads while remaining unrestrained to pursue development without the rigid or formalized overhead that would slow them down.

Client’s Problem: “We have a strategic growth initiative (SGI) aimed at transforming the company from being simply data-driven to being analytics-driven. However, we do not have the leadership experience and personnel (we think) internally to extend our reporting unit into an advanced analytics (AA) group. The large consulting firms we use for most other projects are trying to sell us the Moon and stars, with a price tag that fits. To make things more complex, we would like some in-flight projects to evolve with the AA SGI and include machine learning ideas our existing team has proposed.”

What an imposing problem! A funded advanced analytics effort. We loved this engagement involving equal parts projects, processes, and people. At a high level, the Client’s reporting unit needed a strategic roadmap that aligned them with the advanced analytics initiative. We helped craft that road map and put it into action.

The first step was learning and enunciating the Client’s vision and mission regarding analytics, which involved interviewing the C-level executive team and various department heads with vested interests in analytics. Further interviews with the people in the reporting unit yielded a list of core values they expected in themselves, their mission, and their leaders.

We conducted a thorough self-audit of the existing teams (effort-value of existing projects, a group S.W.O.T., a company P.E.S.T., readiness, competitive advantages). With these inputs in hand, they were synthesized into strategic focuses for the new AA group. Alongside the leaders of the reporting unit, we crafted near- and long-term objectives for the new group to pursue. We then sliced each near-term objective into goals with defined metrics/KPIs and estimated timelines.

A significant part of the road map was people and their development. The short-term solution was embedding Primordial teams to do project work while also assisting in hiring experienced people to lead the new AA group. Another major hurdle was the deployment of new technologies and development tools (e.g., AWS, Docker, GitHub, Airflow, Presto) to create efficiencies for the new group. The Primordial teams owned the roll-out and training on these systems for seamless handover.

Client’s Problem: “Our company is undergoing a transformation via a re-org to better focus on the development of our people. We do not have a legacy of strong technical leadership and thus are uncertain of how to create a career pathing that is appealing to our younger tech employees. This is a critical need for us now and going forward as it impacts hiring and retention of people working on high-value technology projects.”

Our Client was within range of several tech hotbeds and had the added stress of competing against those markets. Unfortunately, providing extensive “perks” inside the office, ala Silicon Valley, was not an option. So, the Client needed to be creative. We were a good fit for this engagement based on our history of growing up within the industry and having some ideas that did not require getting into an arms race with other companies.

Our initial interviews and scoping with employees in IT and HR revealed survey data that painted a clear picture: the technical people wanted to feel a sense of purpose, to be motivated by their work, and to have their work make a difference to the business. These are typical desires for many people in the workplace, so we set to the task of combining these desires with a career trajectory.

The first order of business was to rewrite job descriptions based on industry best practices and standards. We also created new titles, duties, and responsibilities that reflected the kind of work people were doing, going to be doing and wanted to do. We tiered these jobs in seniority based on a combination of experience, expertise, and merit to capture the upward mobility expected from doing good work driven by passion.

Our industry experience has taught us that segregating technical expertise from business leadership on distinct career tracts is a red herring that demotivates professional people. Thus, we built into the managerial tiering modifiers (manager, supervisors), which could be added to people’s jobs as they grew into and desired those responsibilities. The result was lengthy but specific and valued, with titles such as Sr. Data Architect, Manager or Principal Data Scientist, and Director.

Having a career path mapped out did not address the job’s desired purpose, motivation, and difference-making. Our career model was given life by including coaching (bought) and development (built) processes that enforced a plum tree paradigm where management existed only to support. As a result, the people were empowered to manage the work, and the managers led the professional and technical development of the people. This inversion encouraged everyone to self-motivate because the individual now determined their career path and work purpose.