Automating field tests to boost operational efficiency


      Previously published on Light Reading

      Question: How many field technicians does it take to change a light bulb? 
      Answer: Two -- one to change the bulb, and another to enter the test data.

      While that gag may not get me a gig at the Montreal International Comedy festival, it will resonate with telecom field technicians who struggle with legacy test and measurement tools. Telecom field technicians, many of whom are now outsourced, spend as much as half their work time performing "bookkeeping" activities such as entering data into test units and uploading test results into databases. And yes, in some cases two technicians are actually sent to carry out a test: one to do the test and the other to enter data. No joke.

      Often the data inputted by field technicians contains errors and needs "scrubbing." For example, the wrong test-point IDs and test-result filenames might have been entered. Incorrect test steps and individual tests may have been performed. Another common problem is that the wrong pass/fail thresholds may have been used for a test resulting in a false positive or negative. Often the mistakes can just be down to simple data entry mistakes such as typos or forgetting to fill in all the fields. The upshot of all this is that field technicians waste excessive time on tests that generate low-quality data.

      Further up the chain of command, we have the project manager who receives reports (closeout packages) from the field technicians. These reports, containing thousands of test results, are often in PDF format, which is great for human readability but hopeless for analyzing the enormous quantity of data involved. Consequently, most project managers are only able to conduct spot checks (verification) of the test results for compliance with network standards. Obviously, this leaves plenty of scope for non-compliance and resultant network quality problems. Needless to say, this can lead to service issues, delays in construction and deployment, cost overruns, subscriber churn and reputational damage for the service provider.

      Telecom operators need to move from complex, error-prone manual test processes, where too much time is wasted on admin, to more automated, consistent and streamlined test procedures where admin is minimized. Automated test procedures (delivered digitally, direct to the tool or smart device) not only make the field technician's life easier, but the higher-quality data they collect can enable managers to better evaluate project activity, network deployment progress and contractor efficiency.

      To make the technician's job simpler (thereby minimizing training requirements), the job definition should include predefined test point IDs (e.g., fiber/bulkhead connector), the specified test (e.g., transport link validation) and context-sensitive, step-by-step instructions. Configurations should be flexible, so that technicians can carry out their tasks without unnecessary delays. Test results should be automatically logged and uploaded (without the need for post-test spreadsheet data entry) to a centralized server.

      This centralized server should have an analytics engine that identifies any detected errors (network standard non-compliance) and requests the relevant remediation (ideally while the technician is still in situ). The consolidated test results of all field technicians might also be analyzed to determine contractor efficiency, network progress and project activity.

      For many operators, today's field testing is highly inefficient, and technician productivity is low. This is affecting their ability to deploy new infrastructure fast enough to meet market demand or identify faults to keep existing customers happy. With test automation, the time to implement tests might be reduced by as much as 50%, and the time wasted on post-test data entry could be eliminated entirely. Furthermore, automation can reduce the error rate of test data both at data collection time and when inputting into systems to pass on to field managers. Reduced human error translates into higher QoE for customers and lower opex for the telcos. And that's no laughing matter.