Bidding To Win Part 2: Situational Bidding Techniques (19)
Bidding To Win Part 2: Situational Bidding Techniques
Welcome to Part 2 of STM's guide to the black art of AM - bidding. Bidding can be the difference between a loser campaign and massive profit - but it's hard to figure out, and there aren't many guides out there. I'm aiming to change that.
Last week we looked at an overall bidding strategy to map out a traffic source. This week, we're going to get more granular.
Bidding isn't something that we can write a single universal guide to. So here are 5 bidding strategies that you can use on specific platforms, at specific times or in specific situations.
Bidding On "Quality Score" Platforms: FB, Adwords, Bing, other Quality Score platforms
Platforms that use some kind of "Quality Score" - Facebook, Bing and Adwords along with a few of the newest ad platforms just emerging - don't respond well to an exploratory bidding strategy, or one that's focused on bidding low.
Why? Because they use the performance of your ad to determine the price you pay per click. And just like on any other traffic source, the performance of your ad will be directly affected by the quality of the traffic your ad sees.
So the best strategy on any of these platforms is usually to bid high - even very high - at the beginning of your campaign. That'll ensure you get the best quality traffic, which will ensure that your ads perform as well as possible. And that will usually mean that the price you pay for a click drops way below the price you'd be paying if you bid lower initially.
There are more - much more - detailed tips and tricks for optimising each of these platforms, but as an overall rule of thumb, this one will take you a long way toward optimised bidding on Quality Score platforms.
Bid Testing For Traffic Quality - Non-Quality Score Platforms
Not all impressions are created equal, even on straight bidding platforms like DSPs (Go2Mobi, Decisive, SiteScout, etc). Generally, if you bid higher within the range of bids for your targeting, you'll get better **quality** traffic - less bots, better placements, and users more likely to convert.
A while ago Tom Fang ran a fascinating study of CPM against conversion rate over on the POF platform. It showed that - approximately speaking - ROI tended to stay the same as bids increased, because quality increased too.
However, that's not true on all platforms and across all targeting. So what do you do?
Once you've mapped out the landscape for your bids, the best thing to do is to create 3 or 4 campaigns targeting a range of bids within the bidding range that will get you decent traffic - one at the top bid, one at about two-thirds up the range, one a third up the range, and one at the bottom of the range. Then let them run for a few days and compare their performance. You'll soon be able to tell if there's a quality difference.
Testing For Time Of Day Changes
Bids can change radically over the day as other advertisers' budgets run out, or day-parting kicks in.
If you have the budget and the time, the best way to map this change out is to run a mapping campaign (as in Part 1) several times in a single day, keeping it paused the rest of the time. I'd recommend about a quarter of the way through the day, half way through the day, and three-quarters of the way through the day, where "day" here means your traffic source's definition of the start and end of their day. (If the traffic source starts its day in your timezone, then test assuming US hours for the day, as that's where most advertisers probably are.)
Beyond that, keep a close eye on the hourly volume of impressions you get. Watch for peaks and troughs. Try duplicating a campaign at a lower bid, one that normally wouldn't get you enough traffic, and see if the traffic suddenly picks up at a certain point in the day.
Once you have this information you can use it to optimise your bidding. For example, if you discover there's a lot of cheap, good-quality inventory late in the day as advertisers' bids are exhausted, you can deliberately bid lower than you otherwise would to capture that cheap inventory and boost your ROI.
Non-Bid-Related Bidding Factors
It's important to remember that on a lot of platforms, your bid is not the only thing that affects the volume of traffic you receive.
For example, on many DSPs, your daily budget will also throttle the number of impressions you get. If you're bidding on a limited pool of traffic, that can often throttle your campaign to the point that you spend a lot less than your daily budget every day, as the DSP's algorithms get their predictions a bit wrong. Try doubling or tripling your daily budget and watch the increase in impressions - sometimes you'll need to have a daily budget of $100 to spend $33/day, for example.
On some platforms, available funds will also throttle your campaigns. In particular, if your available funds are very low but not zero (less than $100), that can result in your available impressions slowing to a crawl.
Finally, do check if the platform you're on has any restrictions on content. On some DSPs, for example, some kinds of ad will be allocated only "non-brand" or low-quality traffic. Ask the platform's support team if this is the case, or ask on STM. If it is the case, finding a way to get your campaign out of the low-quality bracket will often open up an enormous amount of additional, lower-competition/higher-quality inventory.
Use The Source's Data To Double-Check Volume
Most platforms will have indications of how much volume is available in a specific country or a specific targeting combination. If they don't, or if it appears to be out of date, you can and should also ask the traffic source's support team what sort of volume is available for your targeting combination.
You can then use that number to check if there's something reducing the amount of impressions you're getting from where it should be. This is a very useful trick to diagnose issues like daily budget limits, ad quality problems, or even errors on the traffic source's end.
Don't be afraid to get in contact with your traffic source and ask them for advice - whilst they're not completely impartial, it's in their interests to help you spend money with them!
*And that's it for this week! In the third and final part of the series next week, we'll look at "PvP" bidding - sneaky tips and tricks to outsmart, out-bid and out-profit your competition.*
Yet another EPIC series by caurmen! This one's going straight into AMC course resources - having a bidding strategy is so important and something that's not discussed nearly enough.
Thank-you!
Amy
@vortex - look out for part 3 next week! I'm going to be revealing at least one sneaky bidding trick I've never talked about on STM before...
Waiting impatiently for part 3 
Thank you caurmen!
Is part 3 somewhere and I just missed it?
Right here: http://stmforum.com/forum/showthread...he-Competition
Originally Posted by caurmen
@spartanen - I hear you, and am adding detailed guides on optimising bidding on Google and Facebook to my Potential Future Articles list
On some platforms, available funds will also throttle your campaigns. In particular, if your available funds are very low but not zero (less than $100), that can result in your available impressions slowing to a crawl.
Finally, do check if the platform you're on has any restrictions on content. On some DSPs, for example, some kinds of ad will be allocated only "non-brand" or low-quality traffic. Ask the platform's support team if this is the case, or ask on STM. If it is the case, finding a way to get your campaign out of the low-quality bracket will often open up an enormous amount of additional, lower-competition/higher-quality inventory.
Just a quick question: Which TS falls under these 3 pointers?
1. Daily budget throttle
2. Available funds throttle
3. Restrictions the type of landing pages/content
How do you mean? Which TSes fulfill all criteria, or which don't, or which fulfill some?
Originally Posted by caurmen
I meant which of the TS fulfills each of the criteria?
Eg. Which TS will have daily budget throttle?
Which TS will throttle your traffic depending of your available funds?
Which TS will have restrictions based on the type of LP you have?
Phew, big question! There are literally hundreds of traffic sources out there so it's not really possible to answer this question for all of them.
Also, almost every source will have at least some restrictions on what lps you can run: the restrictions just vary.
Which sources in particular are you interested in?
Originally Posted by caurmen
I'm asking in particular for more of the bigger sources:
Zeropark
PopAds
Propeller
Exoclick
PopCash
Plugrush
Adcash
Aha, gotcha.
OK - this is currently lurking at the bottom of another thread. Can you post it as a seperate thread, titled something like "what features do these traffic sources have?", listing those sources and the features you're interested in?
That should help get you quick replies. I know the answers for some of those, but I don't use all of them enough to be certain in all cases.
Alright, thanks for your tips!
I have created another thread >>> https://stmforum.com/forum/showthrea...mobile-TS-have
Should bid optimization be done before we found winning lander/offer or after we have a positive ROI?
Originally Posted by zyggmunt
On the other hand, sometimes it pays to play with the bids early, in case you are not able to get the conversions going... sometimes you just need a different position in the bidding chain to start getting them.
I do it like this : set something like an average bid and start testing LPs and Offers ... if after a decent amount of traffic, there are no or very few conversions, I try to move the bids up to see if the problem isn't with the position... if it doesn't help, I test more Lps and/or offers.
Originally Posted by zyggmunt
One reason is because it's almost never good to have too many moving parts at the same time (i.e. testing offers/landers AND bids). Another reason is because when I test bids, I would cut placements at each bid (either have several cloned camps running simultaneously to hit different placements, and cutting placements for each; or cut placements, raise the bid to hit more placements, cut another batch of placements again, and raise the bid further etc. etc.). Doing so takes a lot of effort and money and time, so I would only do this for a promising offer+lander combination.
Exception: If I'm planning to do some mass-testing in a geo (testing lots of landers and offers), I would sometimes test the bid first using one or more AM-recommended offer that I'm confident will convert, just to find a bid that seems to convert well. Then I would do the lander/offer testing at this bid. Doing so can save money - but I would only go through this trouble if I'm planning on doing massive testing in the geo.
It's all about making sure that the effort+time+budget you put in will be commensurate with the potential rewards you can reap.
Hope that helps!
Amy
Home > Affiliate Marketing Forum >