People, even those that perform analytics, often think that data analysis can only be applied to financial-type audits. I have tried to highlight other types of audits where analytics played a significant role including transportation, inventory, and hazardous materials (environmental). In that vein, I offer you analysis that was part of an HR recruitment audit.
he organization was an international/national police force. Like many police forces, it needed a fairly continuous flow of recruits. The problem with this agency was that the recruitment process – which leads to a six month training program – was overly long. In fact it was 18-22 months from the time a potential recruit entered the process until they were offered begin the training program. During this time, they were not paid, and, as a result, many suitable recruits exited the recruitment process because they found other jobs.
Working with the HR section, the auditors determined that they were 36 separate steps in the recruitment process. Some were fair minor – like completing an application form – while other were more time consuming – like the security clearance process. I was able to obtain the recruitment data for the past 3 years. The data contained the start and end date for each step for each recruit. In reviewing the recruitment data I was able to determine that the steps were done in series – not in parallel. This meant that before a recruit could enter step “n”, step “n-1” had to be completed. Our first recommendation was to change the process to permit steps to be done in parallel. For example, rather than waiting for the results of the written test (which could take up to two weeks), recruits could start on the physical test phase.
One of the first analyses to be performed was to determine how many recruits made it through each step. As expected, over one thousand had completed step 1, but the numbers dwindled dramatically by the time we reached step 36. A second, related, analysis was done to determine how long each step was taking. As mentioned previously, some were quiet short such as completing the application form. By using an expression to calculate the days (end – start) of each step and totalling the days for each step by recruit and then dividing by the number of recruits who completed the step, I was able to calculate the average number of days per step.
The HR specialists were somewhat shocked by the results. Step 28 (a medical exam) was taking over 4 months on average. In turns out that recruits, after having completed step 27, were told that they need to have a medical performed by their family doctor. This was a major problem for those who did not have a family doctor. Even those who did have a family doctor found that the process (calling to make the appointment, going to the appointment, getting the results of the blood work and other tests) took months.
The audit made two simple suggestions: do steps in parallel where possible; and tell recruits that they needed to arrange for a medical checkup at the start of the process so that when they made it to step 28 it wouldn’t take them months to get the results. The adoption of these simple recommendations immediately reduced the recruitment process from 18-22 months down to less than 12 months. As a result, the number of recruits completing the process more than doubled.
We also gave the analysis scripts to HR who continued to refine and improve the process. Within two years the process was reduced to 8-10 months and the number and quality of recruits completing the process improved.
ACL Commands: EXPRESSION; TOTAL; CLASSIFY; and STRATIFY
Lessons-Learned: This was an interesting audit because HR had asked internal audit for assistance. They knew there were problems with the recruitment process – too long and too many recruits were dropping out – but had difficulty dealing with the data. It was fairly easy to look at each step for an individual, but more complex to combine data for thousands of recruits. Since that had asked for our help, we had no problems getting access to the data or having experts help us to understand the recruitment process. They also worked with us to validate the results and continued to use our analysis to monitor and improve the process after the audit was completed.
About six months later, when I needed access to HR data a\for another audit – the work we had done during the recruitment audit helped to grease the skids (made it much easier). I also asked the VP of HR about the recruitment audit and how the ongoing monitoring was going. During our chat she mentioned that audit has been identified as a valuable resource capable of analyzing data to improve the efficiency of operations. This “fact” was well known at the senior management table and it was the VP Procurement who had suggested that she ask for our help.
This solidified two points in my mind. One, personal relationships are critical if you want to continue to work in an organization. You shouldn’t break down doors and hit people with the audit charter statement “audit is entitle to access any and all information” because you will likely be back there again asking for data. It is better to work with the business process owners and system programmers.
Two, when audit consistently performs valued-added work it will be recognized by senior management. The flip side, shoddy analysis, incorrect interpretation, and poor results will also be seen and remembered by senior management. Every analysis has to be carefully planned, executed, reviewed, and verified because you are only as good as your last audit. I am not saying that management is looking for an excuse to shoot down the value of internal audit, but audit needs to not give them any opportunity to do so. We often have more time to examine a business process than the people who are trying to keep it running. We need to use that time wisely and ensure that our recommendations address risks and control weaknesses and improve the business process.