I don't think that's a bad approach - is GDP change specifically provided? It is not what I would instinctively have gone for, but it does make sense.
I agree with Angelo's point about market trends, as I suspect this may impact the total volume more than GDP change. GDP change may show a correlation with spending, but doesn't shed light on what customers are buying. For example, could an increase in GDP mean that more people are going for ground coffee or specialty coffees? With the advent of coffee snobs and hipsters and the relatively low cost of coffee, I would imagine the trend for instant coffee purchase is on the decline, even with GDP growth.
You could also segment your data and make some assumptions about the factors e.g.
Private customer vs companies (e.g. firms providing free coffee for employees, cafes or restaurants using instant coffee)
Generally you'd segment if you thought there was a difference in purchasing habits, so whatever comes to mind is fine.
Hope that's helpful!