Explore Pacvue's Commerce Solutions, including our tech-enabled Revenue Recovery Services.

Learn More

Get Rich or Die Trying: Two Primary Approaches to Projecting your 2020 Amazon Marketing Spend

Budget planning season is open. As Amazon continues to lure digital ad spend away from Google and Facebook and diversify its various ad products, the evolving landscape of online marketing complicates the process of budget planning for 2020. If, like many advertisers, you’re in the midst of preparing for the challenge of projecting your ad spend and strategy for next year, we can help.

To project your 2020 Amazon marketing spend, there are two primary approaches: top-down and bottom-up. Top-down methodology involves using higher level objectives or KPIs along with general heuristics (or rules of thumb) for planning purposes, an example of which is measuring ad spend as a percentage of total retail sales.

These methodologies each have their own applications and limitations. Top-down can serve as a good guideline but doesn’t provide you with much information about what is going on beneath the surface. Bottom-up relies on certain assumptions because of gaps in the data provided by Amazon, so can be more flexible but less precise. In both cases, the fast-changing eCommerce landscape creates uncertainty when projecting into the next year.  

To create the initial plan for your 2020 budget, we recommend starting from the bottom-up and then leveraging a top-down approach to evaluate and fine-tune your projections (we’ll go further into the top-down approach later in the article).

Using a Bottom-Up Approach to Develop Initial Plan

When developing a marketing plan from the bottom-up, the first thing you need to contend with is the limitations of the data that Amazon provides. Amazon provides advertisers with indices on traffic trends but does not provide the precise number of total searches or detail page visits. These figures have to be estimated, so will likely be imprecise. Total category clicks are also not provided, but advertisers can get closer to an accurate figure by using a combination of data points from Amazon and third-party data providers.

Click-Share Model: Your own clicks are easily measurable and the argument can be made that clicks are a good proxy for category traffic. This is the foundation of the click-share model, a forecasting exercise that minimizes the amount of uncertainty by relying on the least ambiguous variables available: your absolute number of clicks and the total category clicks, which can be measured to an adequate confidence level. With this model, your budget projections are based on your category click-share.

Share of Voice Model: An alternative model is the impression-based share of voice. This model is generally less reliable as the quantity of impressions can fluctuate significantly, mainly due to the sheer quantity of on-site advertising real estate available on Amazon. For example, the product detail page can generate 25-50 impressions per page view while the more valuable top of search placements might generate only 2-3 impressions per search; since the latter is a quality-over-quantity placement, bundling both types of impressions together can skew the data. It’s worthwhile to note that most companies would use impression share from the first-page search results to calculate their share of voice, which raises another question on whether a company should use just the top of search to calculate share of voice or also include advertising impressions from the middle and bottom of the page. This is an important distinction since these impressions are not equal in terms of shopper engagement.

While share of voice is measurable, we reason that the data is not as accurate, consistent, or comprehensive as clicks, so the budget planning strategy outlined below is based on the click-share model. Depending on your marketing goals and KPIs, a share of voice model may, in some cases, be a good alternative.

What We Know (Available Data)

  • Current Marketing Spend – Actual YTD spend plus estimated spend through the remainder of the year. This will serve as the baseline from which to project 2020 spend.
  • Clicks – Readily available via the Amazon Marketing performance reports.
  • Cost-Per-Click (CPC) – Based on historical performance from the Amazon Marketing performance reports.
  • Current Market Share – Can be measured by third-party data providers.
  • Historical Category Growth – Sometimes provided by Amazon and can also be calculated from market share data.
  • Retail Sales – Sales forecast for the current year and sales target for next year.

What We Should Assume

  • Click-Share as Market Share – We can assume that in most cases (when ad spend is sufficient and conversion rates are consistent with the category) click-share should be close to actual market share. This is the key assumption in the click-share model, in which a brand’s percent of total category sales is proportionate to total traffic (measured in clicks).
  • Projected Cost-Per-Click (CPC) – Paid Search on Amazon is an auction platform, meaning that brands are bidding against each other to win impressions. Amazon paid search continues to be an attractive area of investment due to its lower-funnel targeting. As such, CPCs can generally be expected to increase over time; brands continue to ramp up investments, which subsequently puts inflationary pressures on CPCs.
  • Click Growth – One can assume that as a category grows, so does the total amount of traffic. This growth is generally proportionate and as such, absolute category growth is a useful proxy for click growth.
  • Efficiency – As a brand enhances its management of paid search, whether through technology or overall better management, we can assume that this is likely to translate an increase in efficiency (better ROAS).

Scenario Planning

Following from the data and assumptions, you can now begin building scenarios for the various investment options available to you. Advertisers will view Amazon differently from a strategic perspective, so these options will vary according to your objectives. For example, one might invest aggressively to maximize market share while another invests conservatively due to profitability constraints or to allocate budget to other strategic channels or retailers. Scenario planning allows you to project the impact of different approaches and choose the option that best aligns with your specific business goals.

Ultimately, the current value of Amazon as a retailer is highly category dependent. No advertiser has limitless resources, so budget planning for 2020 has to account for the strategic value of each channel. eCommerce (particularly Amazon) is a high-growth channel so brands tend to be more willing to over-index their investment, but it’s a valuable exercise to project scenarios and compare the ACoS to your other priority channels.

Additionally, Amazon’s high growth rates may be off-set by the complexity of pricing and the supply chain. Amazon’s packaging guidelines can drive up costs for manufacturers, while their competitive pricing algorithms can reduce profitability for suppliers. Scenario planning can help you quantify all these factors and make an educated decision about your Amazon investment.

The above chart showing a scenario comparison is based on an actual forecast for a Pacvue client. Due to the growth of the category, our data showed that AMS spend would have to increase by a minimum of 35% just to maintain a flat click-share within the category. Further increases in spend improved total clicks significantly but approached a point of diminishing returns as ACoS increased in kind.

Using a Top-Down Approach to Evaluate Projections

When using a bottom-up projection, it is important to benchmark the results. Cross-referencing your spend projections against a top-down framework helps to validate and fine-tune your estimate. This is a key step given the amount of assumptions and third-party data which factor into the click-share model.

This is where ad spend as percent of total retail sales comes in. This is category and brand dependent, but in most cases Pacvue recommends investing 10-15% of total retail sales into paid search to remain competitive. The ratio is calculated by dividing the total spend projected via the click-share model by next year’s sales forecast. If, after calculating your ad budget from the bottom-up, you discover that the top-down (% of total retail sales) falls well outside of this 10-15% range, you might have to re-evaluate your assumptions to see if they are realistic.

Within this range (adjusted based on your category if necessary), the goal is to find the right balance to suit your objectives. If your top-down calculation falls on one extreme end of the range, confirm that it aligns with your marketing goals: if you intend to be aggressive in 2020, 10% is likely not enough; if you intend to be conservative, 15% is likely too much. Investing too conservatively can inhibit reach and limit discoverability, particularly at the top of search where most clicks occur. Investing too aggressively will put downward pressure on portfolio returns. As spend scales, keyword bids will increase, improving visibility but subsequently increasing ACoS, particularly in categories that are dominated by a few key players.

Finally, it’s important to remember that these are projections. Projections necessarily involve some educated guesses, and the more educated, the better. Good data is key here, and while we can offer some good rules of thumb, the best insight into your particular segment of the market is the data from your own historical sales and paid search performance. So when forecasting for 2020, look back on previous years, measure how accurate your past forecasts have been, and fine-tune your assumptions to achieve more realistic results. Updating your tactics in response to actual results is always the best strategy, especially in a volatile, high-growth channel like Amazon.


Awards & Recognitions