With 2018 around the corner, planning is in full swing for next year’s strategies and budgets. It’s a good opportunity to review what elements to keep in mind when developing your projections.
1. Current Trends Aren’t Always a Good Predictor of Future Trends
Many forecasting approaches are based on some modification of what is currently being done.
One may take a paid search campaign, look at how its traffic has grown over the year, and project a similar rate for the near future.
Even if accounting for improvements due to keyword optimizations, copy testing, and similar steps that would change increase efficiency, it is misleading to assume a similar overall trend without considering broader demand factors.
It is worth reconsidering what may affect overall traffic to begin with and if the nature of the traffic itself might change.
2. Look to Non-Digital Insights to Guide Your Online Marketing Planning
In today’s digital world, we rely on our devices like never before.
With our personal experience as a reference, it’s easy to have a digital bias and place a strong emphasis on online data points like online paid search impression share, online form submissions, or website sales.
However, leading indications often can be observed across non-online channels and inform online marketing planning.
Algorithmic forecasting tools, like Google’s Keyword Planner, that rely on digital only insights and infer results from past data wouldn’t incorporate this context.
This is particularly relevant in industries where advance purchase and consideration is key, such as in travel, where users will research decisions months in advance.
Here, getting insights from call center inquiries or sales team who handle RPFs for weddings or large groups, is a good indicator if a destination may be more or less popular than the prior year.
Similarly, in the B2B space where considerations cycles are long and conversions occur offline, strong web traffic and online leads may mask underlining trends of conversion rates that are starting to fall but not apparent if looking solely at online data.
3. Consider Trends in Complementary Industries
Trends in a related industry are also worth watching. The principle of complements suggests that if demand for one product declines, so will demand for another.
A drop in search traffic for “London flights” is a good proxy for future demand for London hotels.
Similarly, a retailer can look at what products are typically bought together or in sequence.
For example, in forecasting future budget for a printer cartridges campaign, it is useful to explore trends for printers.
4. Segment & Use a Bottom-Up Approach
Rather than looking at overall demand, segment it into buckets.
Forecast demand separately for each segment and combine the individual forecasts for a more accurate combined picture.
The longer approach is well worth it. You might uncover patterns that otherwise would have been missed.
If your SEM strategy already incorporates audiences, it’s quite easy to analyze each audience’s performance. You can link set them without implementing bid modifiers or rules, but still get valuable insights on user behavior of each audience segment.
You could well find that modest overall growth is due to the decline of one audience segment offsetting the growth of another segment. Or that visits from net new users are growing faster than referrers from other channels while driving revenue growth at slower rates.
Conclusion
Forecasting paid search, organic search, or any digital channel performance is part science and part art.
Various tools and your search engine support team, if you have one, can help with the quantitative aspect of forecasting.
To put it into perspective, stress test it and adjust it based on the above analysis.
Lastly, track the accuracy of your forecast over time. It is valuable to compare your past forecasts versus actual performance and track the variance over time. This will help refine how to make assumptions around qualitative forecasting aspects.
More Paid Search Resources Here:
