You’re right, the contribution margin, also known as marginal profit per sale (profit will change depending on volume, so it measures the profit received from one extra sale), would be extremely difficult to calculate for retail stores with multiple different products and pricings.

Although the different profits for each product make it almost impossible to carry out accurate pricing analsis, there are workarounds. Calculating average contribution margin, based on an average basket of goods (this can then be updated monthly, for example) or perhaps focusing on a return on investment (ROI) figure could be useful. Remember, with retail, everyone’s in the same boat, so any way you can find to better measure PPC efficiency will help to deliver better results.

Accurate CPC data for every position is only really possible by adjusting bids for high-volume exact-match keywords and looking at positions over time, although this can be subject to temporal bias (certain hours or days of the week can have higher competition and could bias CPCs). Google recently launched a Bid Simulator which shows likely CPCs if you were to raise or lower your bids, but since it’s only available on a per-keyword basis, it doesn’t really help for large profit maximisation calculations.

In terms of Andreas’ calculations, they all seem fine to me, and his conclusion that profit sharing is a superior model to revenue sharing is spot on. However, with any pricing model, it’s important to find one which works best for your business. Whether profit sharing is your ideal model will depend on many factors such as exisiting relationship with the client, level of trust, amount of keyword cost and revenue data, confidence in conversion rate data and strength of brand.

Only you are able to decide which pricing models will work best for your situation, but hopefully this article, and the last two on markup pricing and performance pricing, will help you arrive at a more informed decision.

Good luck!

Alan

Dina Carroll…

I just found your website and it is a great reading. I am looking into profit sharing after I was reading this article:

http://www.vinnylingham.com/specialreports/profit-sharing.html

It is all very interesting but I had some questions regarding this article and maybe you are able to answer them.

- How do calculate the CM (Contribution Margin) when somebody has a shop and promote all kind of different products in ppc ?

- In this model they use cpc but how do know the cpc for every position?

- Do you think the above article is realistic with his calculations?

Any advice would be great because I need to get my head around this.

Thanks

Menno

Renate Verbaan…

Kimberly Blair…

I would tend to advise against revenue sharing. Since revenue sharing does not take into consideration click costs, it does not create incentives for the agency to spend effciently.

If you’re taking about profit share, which does takes into account PPC costs, the percentage really depends on many factors such as relative bargaining strengths, reputations and likely risks of PPC activity for client and agency, average profit margins for the client’s products and services, the competitiveness of the industry and the estimated work, effort, knowledge and expertise required to make the PPC campaigns a success.

Thoughts? No attribution worries.

This post was mentioned on Twitter by alanmitchell: Economics of PPC Pricing: Why Profit Sharing is the Future http://bit.ly/cGNOiR #PPC #Economics…

Good point about assigning true value to brand vs. content vs. generic keywords. Different types of keywords naturally require very different levels of effort to achieve profitability, so it makes sense that agency fees are rewarded based on actual efforts. Brand vs. non-brand performance pricing is something I touched on in a previous post on cost-per-sale models, although it starts getting incredibly complicated once multiple variables are taken into consideration.

Interesting point too about profit measurement and correctly assigning PPC contribution per order. I guess it depends on how open the client is willing to be regarding their profit margins for each product, as well as the number of outside influences with might affect sales volume, such as promotional discounts and non-PPC marketing activity. I think the smaller and more simplistic the client’s business model, the more likely a profit sharing model could work.

@ Tom,

You’re right, short-termism is definitely a big problem with any performace-based model. Using engagement metrics such as conversion funnel goals or on-site behaviour would help remove the problem of short-term bias, as would separating budgets for brand vs. non-brand, but a longer contract length could be a more flexible and efficient solution if a good relationship already exists.

If a contract was to last for only a few months, the agency would see little point in developing the client’s brand for long-term profitability, so would only focus efforts on easy, high-converting keywords. However, if a contract was agreed to last for multiple years, it would then make sense for the agency to take on the client’s business interests as their own, and work to develop long-term profitability.

Excellent point too about the length of the purchase cycle. The shorter the time from research to purchase, the easier it is to accurately attribute value, and the better the model becomes at accurately reflecting the efforts of the agency. Big multinationals and B2B suppliers should probably use the profit sharing model with caution.

I agree in that this model can promote a degree of short-termism in the agency approach which does not necessarily take into account longer term strategic objectives – i.e. brand building and development.

Some solutions to these problems are to a) reward effort based on an engagement metric (i.e. reward a conversion path touch for all completed actions) and b) ensure that there is a separate budget for investment in ‘brand building’ keywords/tactics.

Finally, worth noting that this model tends to work best when there is a short research/consideration/purchase cycle – i.e. there is low latency between research and sale. As the research period gets longer it becomes harder to attribute value (based on cookie length/deletion, complexity of conversion path).

However, don’t think even profit share is the panacea to the agency pricing model. You mention attribution as one, others could be;

Brand / Non-brand split – should I be giving away the same level of profit for brand attributed searches (no real work required.) Or product based keywords? Or content network attributed sales? Should I be giving most away for generic campaigns (hardest to make profitable)?

Profit measurement – my big thing is measuring true contribution per order. Is profit going to be just on product margin or take into account all variable costs? Should it be averaged, I might make massively different margins on different products. I might have a discount code in the market at the time which could be the true driver of conversion rather than an efficient PPC campaign.

Profit as KPI – I’ll know that some keywords won’t be profit making but I might want to rank highly to gain the benefit of the impressions on brand awareness. The model would have to be able to take this into account.