SEM Profit Maximization

September 4, 2013

SEM is unlike other marketing channels because you have the luxury of choosing your profit margins (by setting how much you are willing to pay for a sale) and you are effectively locked into the resulting customer volume. This is a very powerful aspect of SEM but one that requires some sensitivity analysis to ensure you are maximizing total profit.

Let’s say you sell premium WordPress themes online, and you sell each one for $30. Let’s assume you own the themes outright and there is no other direct cost involved with each incremental sale. So you can spend up to $30 to acquire each customer, but what should you be paying to ensure you make the most money? A lot of marketers fall into the trap of being complacent with an arbitrary profit margin, not knowing how much they might be leaving on the table. You could come to the conclusion that you want a 50% profit margin and so you are willing to pay a $15 CPA (cost per acquisition, a common term used in online marketing). Let’s say that at a $15 CPA you can acquire 500 customers each month. So each month you spend $7,500 to generate $15,000 in revenue and you are left with $7,500 in profit. This might sound great but you are ignoring many other combinations of profit margin and total profit, some of which will likely result in more total profit.


The Total Profit Curve

It requires just a little bit of math to understand how to maximize profit when using SEM. Total profit is simply total revenue minus total cost. If you were to graph total profit, it would be a concave parabola that has some “peak” which corresponds to the sales volume where profit is maximized. Not to dive too deeply into economics, but the reasoning for the shape of the total profit curve is because as you seek more sales volume your costs will increase because you need to “work harder” to find more customers and that causes an increase in cost and thus a decrease in profitability. This could be hiring more expensive employees, using more expensive marketing channels, etc. For the purpose of SEM, the reason for the increase in costs as volume increases is due to keyword pricing. To get more volume you have to bid more for keywords (which is equivalent to lowering your profit margin) in order for your ad to show up more frequently or in a higher position for those keywords. Below is an example of what the total profit curve might look like for our hypothetical WordPress theme business.


Graph of Total Revenue, Total Cost, and Total Profit

Graph of Total Profit Curve


Profit Margin vs. Total Profit

Our total profit formula [Total Profit = Total Revenue – Total Cost] can be further written out as [Total Profit = Sales Volume * Revenue Per Sale * Profit Margin]. If you notice, you have complete control over all of the variables except sales volume, which is partially determined by you from your desired profit margin (effectively your keyword bids) and partially determined by Google’s keyword auction system that allocates you traffic based on your desired profit margin. And that’s the point of leverage. You can experiment with different profit margins and discover the associated sales volumes and resulting total profit.

From above, we know that when revenue per sale is $30 and profit margin is 50%, sales volume is 500 customers per month and total revenue is $7500. What we don’t know is how customer volume changes when we adjust profit margin. So what we have to do is adjust profit margin and take note of the effect on volume. For example, if we hold revenue per sale constant and changed the profit margin to 66.67% (we change our target CPA to $10), let’s say sales volume becomes 200 customers per month and so total profit becomes $4000 per month. That’s clearly undesirable because you just lost $3500 per month in profit. Let’s move the other way and reduce profit margin to 33.33% (target CPA now becomes $20) and let’s say this results in sales volume of 1000 customers per month and so you make $10000 in total profit. Note that this is better than our initial total profit, even though the margins are smaller. After you test various CPA targets, you will have a gauge of what kind of volume to expect and you can build a model like the simple one I’ve laid out below. You can see that of the five CPA targets, $20 is the one that results in the most total profit.

Total Profit Model


Easier Said Than Done

Testing profit margin and volume combinations is important because it helps you understand the total profit curve and where you should be positioning yourself, but like anything, this is easier said than done. While it won’t be better than your own observations from testing, Google AdWords has a really helpful tool called Bid Simulator that predicts click volume for a given keyword and bid input. So if you have solid conversion history you should be able to extrapolate what kind of sales volume you can expect at various bid levels.


Google AdWords Bid Simulator

Screenshot of Google AdWords Bid Simulator

It takes a lot of time, patience, and money to get the necessary data to understand how your target CPA will affect your sales volume, but it’s well worth it if your goal is to maximize profit. You might ask why anyone would not want to maximize profit, but you’d be surprised how many companies (big ones, particularly) don’t look at total profit as the most important metric when utilizing SEM for customer acquisition. That’s a topic for another post :)

Clarifying note: Desired profit margins determine CPA targets:  [Target CPA = (1-profit margin) * Revenue Per Sale] which determine keyword bids: [Keyword Bid = Target CPA * Keyword Conversion Rate].

Using Facebook and LinkedIn Ads to Intelligently “Poach” Employees

September 18, 2012

Recruiting top talent for your startup is a huge challenge. In fact, it’s probably the biggest challenge for most companies. There are only so many awesome people to hire and Craigslist postings and referral bonuses only go so far. So you have to be creative to get above all of the noise. One really cool example is Full Contact’s paid paid vacation policy. It got them tons of press and probably caught the attention of a bunch of talented people they wouldn’t have otherwise been able to get in front of. Another clever technique is to go to LinkedIn and send a message to talented people who are currently employed. Putting aside the questionability of poaching employees of other companies, since you are contacting people who aren’t necessarily actively looking for a job, your conversion rate (interested candidates) will probably be pretty low and it doesn’t scale very well so you might not end up with a sufficient pool of candidates. A good marketer knows conversion rate and conversion volume are inversely related, and the cold message approach doesn’t lend itself to either.

What would be better is if you could target many relevant people working at specific companies (competitors, companies that only hire top talent, etc.) who might entertain a new career opportunity should they see something that interests them. And you can do just this using Facebook ads and LinkedIn ads. Both of these ad platforms allow you to target at the company level, so you can setup a campaigns and show ads only to people who work at specific companies. This allows you to get much more reach than individually contacting people, and interest is initiated by the potential employee so your conversion rate will be higher.

Here’s a real world example: we were recently looking to hire an Account Executive to join us at Circl. We posted on Craigslist and we tapped our networks and we got a lot of talented candidates. But let’s say we wanted to expand our candidate pool even further and we knew there were a lot of companies in our space that might have relevant candidates. I could have created an ad campaign on Facebook that might look something like this:

Facebook has a huge reach and people use it very regularly so your potential audience pool is in the hundreds of millions. But as you can see, the number of users this ad targets is very small (340). Granular targeting is the key to facebook advertising (and any advertising really) so you can create relevant ads that lead to relevant products/services/destinations for a specific audience. Since this audience is so targeted, you would want to bid to the higher end of the suggested bid range and potentially 2-3x the highest suggested bid. The reason being you want your ad to show as often as possible since your campaign is so targeted and you want to reach as many of these 340 people as possible.

We could have also launched a LinkedIn ad campaign. LinkedIn users tend to be more business oriented and in a professional state of mind than users on Facebook. That’s why you see most advertisers on LinkedIn advertising business-related products/services. Seems like a great place to advertise an open position, right? So we could have created a campaign with targeting that might look something like this:

One thing you have to be mindful of when advertising on LinkedIn is that if your target audience has less than 1,000 people, they won’t let you continue setting up your campaign. This is to ensure sure enough people see/click on your ad so you get some results (and they make their money). Fair enough, but this affects the ability to target at a super granular level so keep it in mind if you start targeting smaller companies and/or specific job functions.

I hope this recruiting hack helps you land a new employee or twelve.

Conversion Tracking for Facebook Ads

August 17, 2012

Conversion tracking is a crucial part of advertising. If you can’t track what’s working, you won’t know what to focus on. For some reason Facebook doesn’t fully agree, as there is no native conversion tracking for Facebook ads. There was at one point. Then there wasn’t. Then there was for certain advertisers. And that’s where we are today. You cannot have a successful ad platform without letting advertisers accurately measure performance, so you would think Facebook would want to lower the barrier to get as many advertisers as possible on board. Until that changes, you can use non-Facebook tools to track conversions.

If you have a large monthly budget, it can make sense to use a third-party ad management tool to easier manage your Facebook ad campaigns. There a growing number of these tools and they usually have built-in conversion tracking. But these tools can be expensive and they are definitely not for everyone. Another way to track conversions is to use Google Analytics combined with UTM URL parameters. It’s straight forward and doesn’t cost anything but a few minutes of time to setup. First, make sure whatever site you’re sending traffic to is setup on Google Analytics. Then create a goal within Google Analytics for whatever your conversion is. Make sure your goal URL is the post-conversion step (i.e. thank you page, order confirmation, etc.) in your funnel. Now create a UTM URL for each ad, changing the parameters accordingly so you can segment your conversions by different ad copy, image, targeting, etc. And that’s it. Now when you send traffic to each of these UTM URLs, you’ll be able to use Google Analytics to track conversions by the various UTM parameters you used. This also works great for other traffic sources where you can’t directly track conversions, such as email campaigns and social media.