← Return to Blog

How can we improve forecasting for call-centres?

Time is money – and when the goal is to drive cost efficiency, every minute counts. Media cost is always a key consideration in planning, but additional business costs such as call-centre time can have a big impact on ROI. Therefore it is imperative to think outside of the media box to maximise efficiency across a business’s operations. One way of doing so is to provide accurate call forecasts to call-centres, which in turn can;

  1. Minimum wasted time, reducing costs.
  2. Maximise answer rates, increasing successful responses.

Forecasting response is a key part of what we do, but forecasting for TV brings its own set of challenges, one of them being that programming (particularly when live) can be unpredictable and break and spot times change from those that were forecasted.

When providing response forecasts to call-centres, spots in live sports or around the news are particularly prone to change, which can lead to the problem of call-centre agents sitting on their hands waiting for calls, or worse still not being prepared for large influxes and potentially missing out on sales.

We had a deep-dive into the challenges of forecasting spot times and to explore more ways in which we can improve our process for a charity client in the Netherlands.

What we found

  • The mean average difference between spots forecasted and actual time was 2 minutes later than forecast – but 1 minute later than forecasted was the mode.
  • Reassuringly, only 2% of the spots had a difference of greater than 15 minutes.
  • The majority of spots with a difference in time were aired later than forecast.
  • Looking at the outliers – the difference was much higher than the average, with the biggest difference at almost 3 hours! Unsurprisingly the spots with the largest difference were live sports, with spots in snooker and football the biggest culprits. Only <1% of spots in non-sports programming aired later than 15 minutes after the forecasted time.
  • Looking at peak time only, the average variation was greater, with spots airing on average 4 minutes post estimated time. However, the only spot with a difference of >5 minutes was after news programming.


All Response Media viewpoint

Cost efficiency and maximum response are at the heart of what we do, so we should be consistently identifying areas where we can improve results; working more effectively with call-centres is one area we can do this.

So how can we apply these learnings and what is best practice when forecasting responses for call-centres?

  • Flag spots e.g live sports or post-news prime spots with a potentially higher variation to the call-centre to better prepare the agents.
  • Continue to monitor trends and identify stations that frequently have higher variations.
  • Share access to LivePulse, within ARMalytics®, with call-centres so they can see their website visits in real-time alongside a solid indication of when spots are aired.

Communication and knowledge sharing are the cornerstones of client-agency relationships, and this example is a particularly pertinent one as it relates to real-time workforce efficiency and aligning an advertiser with their advertising to that end.