Facebook Bidding: How to Rapidly Optimize Campaigns

A lot of the information I have posted so far has been about getting started: setting up tracking, writing ads, building out the initial scaffolding for a properly structured campaign, and so on. But setting up a campaign initially is the easy part. The real challenge comes later, when you are tasked with actually optimizing the campaign towards profitability, managing the bids, scaling up, and actually making the numbers work in a consistent manner.So let’s talk about the next stage of building profitable paid traffic campaigns. I’m going to use Facebook as an example, because managing bids aggressively is crucial for a successful Facebook campaign, but you can easily apply these high-level bidding strategies to any bid-based traffic source. I’m going to go through the exact process I used to rapidly build campaigns from nothing to profitably spending $5000 a day on Facebook.

Start Off Losing Money and Paying For Data

Even the most experienced marketer’s campaigns will not be instantly profitable. In fact, I’ve heard that over 90% of PPC campaigns start off losing money. That means you have to launch 10 diverse campaigns, on average, before you get profitable.The truth about Internet marketing that many beginners seem unwilling to confront is that, when launching a campaign, you shouldn’t expect to get profitable instantly, or even soon. If you do launch a campaign that’s making money right away, consider it an incredibly fortuitous aberration, like a site going incredibly viral. When launching a campaign, remember that you’re not paying for a scalable acquisition channel. You’re only paying for data about what actually works in the marketplace. So don’t be afraid to bid high. It’s much better(and more cost-effective in the long run) to bid high and get enough traffic for statistically significant results than bid low and wait a month before you have gathered enough data to start optimizing. Start with the objective of getting as much data as possible as quickly as possible. On Facebook this generally means bidding $0.30-$0.60 CPC for mot demographics, although you may need to bid closer to $1.00 CPC for the most valuable users (like middle aged women).

Cut Low Performing Ads Quickly, Invest in High Performing Ads

After choosing a broad enough demographic and biddng high enough to get a decent clicks, you can begin optimizing. But don’t be too eager to optimize; if you kill an ad right away before you’ve collected enough data, you’re basing your decisions on the whims of random chance. Be sure to wait until you get statistically significant results before deciding if an ad is effective.Wait until an ad has about 30 clicks before deciding what to do with it. Bidding higher and targeting larger demographics will help you get to 30 clicks and make a decision faster. Yes, this type of extensive testing will cost you more money, especially if you’re testing many ads and targeting options and need 30 clicks on each variation. Sorry; You gotta pay to play.Fear of spending money testing can paralyze you. Think of the opportunity cost of not running profitable ads as soon as possible. The faster you can start running profitable ads, the more money you will make in the long term. If a particular headline or demographic is getting a good CTR, keep spending money on it, even if it’s not converting profitably immediately. Facebook will begin discounting your bids and giving you cheaper traffic soon enough, as long as you can keep your CTR up.On the other hand, if an ad has a low CTR or isn’t performing well, don’t try to keep it alive artificially by bidding high. A good CTR to aim for on Facebook is 0.1%. Cut your losses, kill it mercilessly, and move on. There are so many factors you can test. No need to keep throwing good money after bad.It’s all about CTR on Facebook. Get a good CTR, and the cheap clicks will follow.Expect your CTR to slip over time due to saturation and banner blindness. When you see your CTR declining steadily, don’t bid higher to compensate. Instead, create a similar ad with a slightly different image and start fresh. Sometimes, even reusing the same image with a different color border added is enough to maintain CTR.

Find Gold Nuggets with Reports

Once you’ve gotten a decent amount of clicks and pruned the terrible CTR ads, you can begin aggresively optimizing your targeting. You’ve probably heard the adage that 50% of the money you spend on advertising is wasted; you just don’t know which half.Well, with Facebok Ads, you can find out where your ad dollars are being wasted. If you’ve been getting traffic for a few weeks, and maybe getting close to breaking even on some ads, it’s time to really dive into Facebook’s click reports. You can find the “Reports” link on the left sidebar in the Facebook Ads UI. I usually find the Responder Demographics report most helpful- it will break out exactly which users are giving you the most clicks and best CTR (and the cheapest impressions). Double down on targeting them, while still keeping an eye on conversions.If you’ve started by targeting a broad demographic to get enough data, now is the time to whittle down your targeting to tightly focused ads(i.e targeting a specific age, location, or interest) based on those reports.As always, be wary of making decisions based on statistically insignificant data. Use Facebook’s reports combined with your own analytics to figure out who your most valuable clicks are coming from.

See How Low You Can Bid

Ads that have a few weeks of solid performance history are the perfect target for aggressively optimizng your bids. Don’t just settle for the amount Facebook is bidding you down to. Bid lower, wait a few hours, and see if you’re still getting impressions. Keep lowering your bids for ads that have already been running succesfully until you stop getting impressions. You now know your bid floor at that CTR. You’ll be surprised at how low it can be.You can often bid 40-70% lower than Facebook’s suggested bids without losing a lot of traffic.Even in 2011, it’s still very possible to pay pennies a click and get good volume on Facebook- as long as your ads and targeting are strong.

Scale and Replicate

After a few weeks of constant tracking and optimization, you’ve probably got a campaign that is at least slightly profitable. You’re finally ready to start scaling it up and really realizing the benefits of the vast amount of traffic Facebook can offer. Because you’ve already identified the most effective ads and targeting options, scaling up is a lot less risky. You’re already entrenched, established. If you can monetize international traffic, translate your ads and build campaigns for other countries. You’ll find lots of incredibly cheap traffic there if you know where to look.Or, soften your targeting. Instead of targeting 18-25 year olds, test your strongest ads targeting 17 year olds. Use keyword research strategies to find other related keywords or interests. Check your referrers and see which apps/pages your clicks are coming from. If people are disproportionately clicking your ads when using a certain app/game, maybe you can negotiate a direct ad buy with the developers.

  • Confettigirl

    Thank you so much for the information!! What’s your opinion on sending traffic directly to a sales page (physical product ecommerce site) vs to a facebook page for sales conversion? Thanks!!

  • Goldstein Dan

    Thanks! What do you use on the backend to track your conversions – would you recommend using prospers demographic variables?

  • Ilya Lichtenstein

    I pass in a unique subid in the URL of every ad that identifies the demographics. For example, &sub=13m_us, etc.

  • Ilya Lichtenstein

    If you’re interested in likes/getting more free social traffic, sending traffic to a Facebook page can be effective. The best approach is probably to offer a coupon on your Facebook page and link to your sales page from there.

  • JP

    Excellent post as always, lots of nuggets of wisdom that I’ve written down.Thanks!

  • http://hackfacebookaccounts.yolasite.com/ hack facebook free app

    Amazing loads of superb material!