Jobscience provides recruitment marketing, recruiting and contingent workforce management systems for agencies and direct employers in thirty-two languages worldwide. In this article, I would like to summarize what we expect to accomplish with a major project that is underway. Our company is working on a forecasting program that uses Key Performance Indicators (KPIs) to provide guidance in achieving our customer’s goals.
The project is based on mining data produced by the daily operating activities of tens of thousands of users in our system. The system provides interactive dashboards for users define goals and monitors the likelihood that their goals will be achieved. The system employs statistical analysis to identify patterns in the KPI data that are factors in achieving the goals. Artificial Intelligence (AI) adaptive learning techniques are being applied to improve forecasting. The key points are that the KPI Forecasting system relies on clearly defined goals, plans and massive amounts of data to forecast the probability that a user will achieve their goals without having to take corrective action.
It should be helpful to use a hypothetical example of a recruiter that is an employee in a 20-person recruitment agency in the United States with 18 people using the Jobscience system for business development, order fulfillment, shift scheduling and billing temporary contractors. The recruiter works a “full desk” in that 50% of the fees earned for the agency by the recruiter are for providing permanent placement services and 50% of the fees are from providing temporary contractors. The recruiter has a goal of making $75,000 in the upcoming year, including 15% in commissions and bonuses from both temporary and permanent placement fees.
The goal is clear. The recruiter wants to make $75,000 and that requires the right combination of four KPIs: (1) The number of permanent placements for the year, (2) The average fees earned per permanent placement, (3) The number of temporary placements for the year and (4) The average fees earned per temporary placement. These four factors directly impact the recruiter’s compensation goal and, depending upon how they trend compared to past performance, we can forecast the likelihood that the recruiter is likely to achieve their goal.
There are many actions the recruiter might take to improve on past performance. In this example, assume that the agency’s owner wants to focus on temporary placements for the agency to grow faster. Different tactics can be pursued to boost the recruiter’s productivity within the temporary placement segment of the business and KPI measures can be established to measure progress against the plan for the recruiter. Customer contacts, new accounts, job orders, candidate contacts and applications are examples of KPIs to assure that the recruiter stays on track with their plans.
The Jobscience KPI Forecasting system aggregates user productivity data by type of service, business unit, branch location and the entire company to forecast the probability of achieving the goals set for each business entity. Let us return to the hypothetical 20-person agency to demonstrate how this works at a level above the individual recruiter.
Assume that our hypothetical agency currently employs 10 recruiters, 3 sales reps, 3 customer support people, 3 administrative people and the owner. (There are 13 income producers, 7 management and support people.) Fee income from temporary placement is growing 3 times faster than the fee income from permanent placement so at least 3 more people in customer support and administration, along with a dedicated industry sales rep, will be required in the upcoming year based on expected growth.
Further assume that the owner wants to sell the agency and retire within 3 years if the agency is worth at least $5M by that time. (The owner thinks the current enterprise value of the agency is worth on the order of $3.5M to $4M.) Total revenue last year was $10M ($8.75M from temporary and 1.25M from permanent) with each contributing about $1.25M in net fee income ($2.5M in total) to the agency. After expenses, the agency produced about $350,000 in operating income last year.
With the owner’s goal is to sell the agency in 3 years if it is worth at least $5M, the owner needs a plan to create an enterprise worth $5M within 3 years. The owner concludes that the agency cannot achieve this goal if the agency stays on course and operates as it has in the past. The agency has to grow its annual fee income and operating income by at least 15% a year within 3 years.
Last year, revenue from temporary placement grew 14%; fee income grew by 17% and operating income grew by 16%. Revenue from permanent placement grew 6%; fee income grew by 5% and operating income grew by 4%. Combined revenue grew by 10%; fee income grew by 11% and operating income grew by 9%. Significant changes are required to reach the 15% goals for both fee income and operating income in 3 years.
The owner decides that reaching the fee income 15% growth rate goal requires changing the business mix from 50:50 temp and perm last year to 60:40 next year, 70:30 the following year to 80:20 by year 3. The owner also decides to focus on placing contract professionals where the fees and the utilization rates are higher.
Moving operating income growth rates up from 9% to 15% in three years will be difficult. Increasing employee productivity will increase operating income growth rates, however, the shift in the business mix to more fee income from temporary placement requires hiring more customer support and administrative personnel (lowering the income producer share of staff expenses and employee productivity).
Assume the owner decides to increase the agency’s focus on hospital nursing shifts and to hire a sales rep with experience selling to nurse administration. This is expected to increase both fee income and employee productivity by placing more contract professionals for longer periods.
Overall, the goals for the agency are now clear. The goals are to reach the 15% annual average growth within 3 years and have a business that is worth $5M. The agency’s plan for achieving these goals are to change the business mix to capturing more temporary placements for contract professionals targeting healthcare. Goals and hiring plans for implementing this plan can now be developed for each business unit through a cascading planning process.
Through these examples, it is clear that the KPI Forecasting system must provide a “plan builder” for each level of a user in their organization. Jobscience has developed interactive graphic widgets to provide a systematic way of capturing the goals and plan data, as well as providing an easy way to select the appropriate KPIs that each user wants to track. The Jobscience system is capturing individual user activity data associated with each KPI and the Forecasting system mining that data to analyze the probability of achieving each user goal at each level of the organization.
Jobscience can do this by committing to data warehousing on a massive scale. This allows us to apply machine learning and statistical factor analysis to improve forecasting accuracy as more and more data is collected over time. Most recruitment systems involve transactional database applications that do not retain the data required to perform analytics at this level of detail. The KPI Forecasting system relies on data amassed since the user began to use the system so we have a large and growing reservoir of data to improve forecasting accuracy.
In summary, the immediate value of the KPI Forecasting system is that it will connect user productivity data to achieving our customer’s goals at each level of their organization. Key Performance Indicators (KPIs) have historically been used as tools for analyzing past performance. Our KPI Forecasting system will rely on continuously analyzing data from every user on a daily basis to forecast future outcomes so that corrective action can be taken where required.