Bayser AlignerThe Bayser Aligner is a tool dedicated to territory alignment. It is a VB executable that incorporates a free-for-distribution version of MapObjects (ESRI) which means no software is required prior to the installation of the Bayser Aligner. The Bayser Aligner draws upon years of experience of doing territory alignments for pharmaceutical, medical devices, and diagnostics companies, and our expertise in Artificial Intelligence and Operations Research.
First, the Bayser Aligner recognizes and leverages the hierarchical structure of sales forces: territory district region area. Territory, district, regional, and area maps displaying sales, potential, or the alignment index can be summoned at a click of a button. A whole territory can be moved from one district to another at one go. When an account/zip is moved from one territory to another, its district, region, and area are automatically updated.
Second, not all parts of the country need to be revisited in a re-alignment. Indeed, some parts of the country are best left as is because of rep-customer relationships or other historical reasons. The Bayser Aligner captures this notion by associating three states to an account/zip. “Married” means the account/zip is pinned to that rep and should be left untouched. “Single” means the account/zip can go to any rep including its current owner. “Divorced” means the account/zip can go to any rep except the current owner. The need for the “divorced” case came from a re-alignment situation where a female rep was sexually harassed by a physician in a large account. The company could not afford dropping the account and flew a male rep from a neighbor territory to cover just that account. By declaring the account “divorced”, this ensures the account will not assigned to that female rep
As one would expect, the Bayser Aligner uses a layered structure (states, counties, zips, cities, highways, accounts, rep location, etc.) that can be turned on/off at will. It supports identification, information, zoom/pan, and cosmetic functions such as labeling & color changes. As for the core alignment function, which is swapping of an account/zip from one rep to another, before/after territory workload balance charts are displayed to help gauge the relevance of the tentative change, thereby easing the accept/reject decision.
In addition to its user-friendly interactive environment, the Bayser Aligner offers two handy automatic features. The first one addresses the rep location question: “What are the best locations for adding 10 more reps?” or “ What are the 5 territories that need to be folded in?”. The Bayser Aligner uses an electrostatic charge model where reps are represented as negative charges and accounts/zips as positive charges. Adding reps amount to identifying locations where the attraction is the highest (i.e. largest accessible opportunity). Interestingly, this ideal rep location finder has other uses. It can help assess the optimality of rep locations or suggest relocations. This is accomplished by adding say 10 reps and then eliminating 10 reps. Discrepancies between the add and delete lists capture the insights we are after. Another practical use of the rep location finder stems from the fact that the attractiveness of a rep location is quantified. The relative magnitudes of the top 10 candidate adds may suggest knocking down by 2 heads the 8 originally asked, or, conversely, suggest bumping up the 6 reps asked to 9.
The second automatic feature is the account/zip assignment. This feature is useful both when building an alignment from scratch or when improving a current alignment (realignment). From an operational standpoint, what makes the alignment process difficult is the fact that an account can be assigned to several reps. To that end, we developed a conflict resolution mechanism that uses “hunger” and distances of the reps from the account/zip in dispute. The hunger of a rep captures his/her desire to cover accounts and tapers off as the rep takes on more accounts. What the conflict resolution mechanism does is to assign the account/zip to the most “deserving” rep.
To improve an existing alignment, we use simulated annealing. In a nutshell, a change is made to the current alignment by assigning an account from one rep to another. The chosen account is typically in the periphery of the donor and recipient reps. The change in the quality of the alignment is evaluated by looking at the workload balance, time size of the territory, and disruption. If the alignment is improved, the change is accepted. If the alignment is not improved, the change is not systematically rejected. Indeed, it is probabilistically accepted with a probability predicated upon the magnitude of the change and something called temperature. The larger the change and the lower the temperature, the less likely the change will be accepted. This feature allows the Bayser Aligner to get out of sub-optimal or local minima, enhancing its ability to find better alignments still.