The ultimate goal of an e-commerce business is to maximize profit, not ROI or cost per conversion.
Truthfully, maximizing both sales and ROI are contradictory goals, and digital marketers should find a compromise between those two targets in order to maximize profits. This article uses detailed formulas to explain how marginal profit optimization is easy to enforce in PPC campaigns.
[Editor's Note: This article expands on the original article that was published in Polish and translated to English.]
Profit-driven strategy has been recently promoted by Google, not only in Think With Google articles, but it’s also reflected in AdWords Qualified Professional tests, where numerous questions require an understanding of the volume vs. efficiency trade-off.
PPC campaigns focus primarily on profitability. The main strategies used are target cost per conversion or target ROI (Return On Investment), sometimes called ROAS (Return On Advertising Spending). The ROI formula normally used for financial investment profitability analysis, is not optimal for analyzing advertising campaigns.
In the future, profit-driven bid management should replace the current “target ROI” or “maximize- sales-within-budget” strategies. Only marginal profit analysis and profit-driven strategy make it possible to find the optimal bids for a PPC advertising campaign.
Marginal profit and cost per conversion
Let’s say that an advertiser, selling their product online, makes $20 profit per transaction (conversion). The advertiser has set a goal of target cost per conversion of $15 for the PPC campaign.
If the campaign results in 1000 conversions, spending $12,000 monthly, then the cost per conversion is equal to $12. Therefore, the advertiser decides to increase the bids and as a result the traffic in a particular campaign. As a result of higher bids, the campaign creates 1300 conversions, spending $19,500, so that the cost per conversion goes up to $15. The revenue grows and the cost per conversion meets the target.
Was the decision to expand the campaign correct?
Let’s calculate the cost of additional conversions: Due to increased bids, the campaign acquired 300 more conversions, and its cost increased by $7,500. This means that the additional 300 conversions were acquired at a cost of $25 per conversion, and each additional conversion brought a $5 loss. The marginal ROI is negative in this case (-20%) and the profit decreases by $1,500.
This example demonstrates clearly why cost per conversion or ROI is not the optimal target for e-commerce campaigns.
ROI of the PCC campaign normally decreases if the bids, clicks and number of conversions increase; therefore maximisation of ROI usually leads to significant limitation of the campaign.
The example campaign above, notwithstanding changes still has a positive ROI = 33.3%, but the marginal ROI is negative (-20%). The volume of clicks and transactions is higher, but profit has decreased. This means that this campaign is overinvested and is losing money compared to the optimum result.
The bids should be increased as long as the profit grows. Thus, the optimum point of the campaign is where the growth stops and the marginal profit (or ROI) is equal to zero:
Marginal profit analysis may look complex, especially in day-to-day campaign management. However, a simple formula can help us to find the optimal point using the KPI available in typical campaign statistics.
We are looking for an indicator showing when the increase of bids not only increases CPC and clicks, but also increases profit.
We assume that:
- The conversion value (Vc) does not change regardless of the ad’s position. An average conversion ought to generate the same income (The value of the shopping basket remains unchanged).
- The conversion rate (CR) does not change. In the higher and the lower ad position, regardless, the percentage of visitors who convert is the same.
The increase of bids is profitable if the increased number of conversions compensates for the higher cost per conversion. The relative decrease of profit per conversion (Pc) should be smaller than the relative increase of the number of conversions (C):
Profit per conversion is equal to the conversion value minus cost per conversion (Cc):
Therefore, the increase of bids is profitable if:
We assumed that the conversion value does not change i.e. dVc = 0:
Now, we multiply both sides of the inequality by:
This value is greater than zero; therefore the sign of inequality remains unchanged (it is true if Vc – Cc > 0, i.e. if the campaign is profitable, but in the case of non-profitable campaigns, an increase of bids never makes sense).
By definition, the conversion cost (Cc) is equal to CPC divided by conversion rate:
and the number of conversions (C) is equal to the number of clicks multiplied by conversion rate:
As we assumed, conversion rate (CR) is a fixed value. Therefore:
The right-hand side of this inequality is the profit per conversion divided by the cost per conversion, i.e. ROI.
The fraction on the left-hand side of the inequality is the relative increase of CPC divided it by the relative increase of clicks. It is the reciprocal of the price elasticity of clicks (E). Therefore, the increase of bids should be profitable if:
If ROI = 1/E, the campaign is operating at the optimum level, otherwise there is space for optimization, because the campaign is under- or overinvested. If the total value of conversions is smaller than the total cost of advertising, the campaign generates an operational loss, but even with positive profit, the campaign can be overinvested.
Profitable advertisers in competitive markets often tend to increase bids and move towards overinvestment areas and realize that something is amiss when the statement from accounting department shows an operational loss. However, they had already started losing money earlier, when the ROI > 1/E formula was no longer fulfilled.
The analysis of marginal profit is the most accurate method of campaign optimization and brings better results than target cost per conversion or target ROI. The presented formula makes marginal profit analysis relatively easy: the bids in the campaign should be increased if the current ROI is greater than the reciprocal of the current price elasticity of clicks. Otherwise, the bids should be decreased.
Understanding price elasticity
ROI is relatively easy to calculate, and majority of e-commerce businesses’ KPI includes this measure. Price elasticity is more difficult to calculate, so it is worth examining its meaning.
The relation of Clicks and CPC is a non-linear function. Generally, this function is increasing (the more clicks, the higher CPC) and usually concave (the more clicks, the faster CPC grows). In the limit, higher CPC does not increase the number of clicks anymore. An example chart of Clicks(CPC) function is shown below:
After conversion to price elasticity function:
The price elasticity can be easily monitored by AdWords experiments:
In the example above, the increase of CPC by 16.3% (from $2.02 to $2.34) caused an increase of clicks by 27% (from 484 to 667). Price elasticity is equal to 1.65 (i.e. 27÷16.3), and its reciprocal value is 0.6. Therefore, in this case, an increase of bids will be profitable if the current ROI is greater than 0.6 (60%).
Also Google AdWords bid simulator indirectly shows price elasticity data:
In the example above, a bid increase by 100% should result in 7040 – 6060 = 980 additional clicks (+13.9%). The cost should increase from $702.33 to $1048.95 and CPC should change from $702.33/6060 = $0.12 to $1048.95/7040 = $0.15 (change by +25%). The price elasticity is 13.9/25 = 0.56.
The bid simulator columns are also available in AdWords interface views, which makes the bulk data download easier.
Very high price elasticity at low CPC is caused by low base. When we increase bids and the offered CPC become equal to or more than “first page bid” values, the ads become visible and acquire few clicks, which compared to nearly zero is a high relative change. Therefore the price elasticity is very high. As soon as the impression share grows to 80-90%, the additional impressions start to have a smaller impact on the click volume than the position of the ad.
Of course, the higher position, the higher the CTR. There are no official statistics, but the practice shows that moving from position 10 to 1 can change the CTR from around 0.5% to 20% or more.
However, the greatest CTR increase occurs when the ad starts being displayed at the top of the search result page. The CTR on “top” is usually 10-20 times higher than with “other” placements.
Examining the figures, we can see that the “other” impressions produce virtually no clicks, compared to “top” placements. CPC grows, and the number of clicks grows rapidly here. Hence, the decrease of price elasticity slows here, and very often the elasticity grows and reaches a local maximum.
The “top vs. other” report is one of the most important indicators of click potential. A high “other” impression share indicates the chance for higher price elasticity. That’s why the bid simulator in Google AdWords shows not only the estimated number of impressions, but also the number of “top” impressions.
For practical use, it is worthwhile to define custom columns, such as “top” and “other” impression share, separately for Google as well as for search partners.
Search partner websites include not only search engines, but also websites such as flight meta search (Skyscanner, Kayak) as well as price comparison and auction websites (eBay). These ads perform closer to display ads, with lower CTR and more unpredictable behaviour. So, if at average position 1.3, ads have a CTR of around 1%, the Google/search partners split should explain why.
Together with ad position and the search impression share indicators, this set of columns can be very helpful in seeking areas of higher price elasticity.
If the impression share and top impression share are above 80%, there is not so much space to grow. The price elasticity is very low, and increasing bids should bring a small increase of clicks and transactions. According to ROI >1/E formula, increasing bids at very low price elasticity makes sense only if the current ROI is very high.
The actual values of price elasticity depend on the structure of the auction and therefore can be different in different geographical regions or industry sectors and change over time. However, our practice shows that in competitive sectors the elasticity is around 1 when the ad starts being displayed in top of the page placements, and then lowers. The ROI > 1/E formula indicates here, that only advertisers with ROI > 100% should fight for higher “top” placements. Otherwise, increasing “top” positions, for example, increasing the average position from 2 to 1.5, should make profits lower despite higher sales.
Lifetime Value of Conversion
It is very important to evaluate the conversion value and ROI properly. A conservative approach, where only the direct income from the conversion is taken into account, may result in a very defensive strategy.
Online marketing not only acquires conversions, but creates branding effects and, consequently increased sales from direct and brand traffic sources. Conversions reported by conversion tracking systems, like Google Analytics, tend to reflect just a part of the online marketing impact.
An online campaign may bring cross-device and cross-browser conversions. Many products have longer average conversion paths. Some clicks “only” assist, while the conversion tracking system attributes the entire value to the last click. Online visitors may convert offline, calling the call centre or visiting stores at physical locations. A new customer may become a loyal client who will bring more value in the future, not only by direct purchases, but also by referrals to potential clients. Finally, if we don’t place our bid in AdWords auction, the prospective buyers can’t find our website in Google. As a result, our competitors take the available space at lower cost.
Although the value of indirect benefits is usually difficult to estimate, the conversion value used for the calculation of campaign targets should also include the value of indirect benefits.
The current, pure online conversion value is only the tip of the iceberg. Focusing on one purchase income and last click conversion misses the bigger picture because it does not represent the entire value of conversion. Using them to calculate the current ROI may result in too conservative bidding and subsequently missed opportunities.