WORDS
When initially setting up your campaigns, DO NOT go to some keyword tool and dump in a gazillion keywords into your adgroups and campaigns– unless you want to boast about how many terms are in your keyword portfolio. The majority of these terms, scored by whatever techniques are going to be junky and low volume– and the engines will penalize you for it. Rather, what I do is hand pick just a few terms per ad group and then borrow ads from competitors that are already bidding on those terms. Sounds simplistic? Well, it is– but it works. Make sure you group your keywords tightly, so they all reflect the same user intent. The engines will choose one of the ads from that ad group to show, so make sure that each keyword is just as relevant for the ad you show.
MONITORING AND OPTIMIZATION
This is probably where most people fall down, because they try to get really fancy and end up spending lots of time optimizing garbage— spending their time in the wrong places. What I do is let a campaign run for a day or two (less if there’s lots of volume) and then look at which terms are driving the most volume— sorting first by clicks descending in AdWords Editor. By the way, if you aren’t using AdWords Editor and are trying to optimize via the web interface, you are handicapped right there because you can’t make bulk changes and do “Excel-like” things like sorting, find/replace, and copying. So I spend just a few minutes looking at what keywords are driving the most clicks and see what the conversions are looking like. The question I get all the time is “How much data do you need to make a decision?” Depends on what the offer payout is. If it’s dating and you’re getting $3, then maybe you’d spend up to $6 on that keyword/ad group to determine whether it’s worth continuing. Or maybe it’s a small biz credit card offer that pays out $150— you’d be willing to spend a lot more. There is a statistical significance formula for calculating sample size that’s based on your expected conversion rate, confidence interval, and threshold for detection– but you don’t need that level of precision here. At this point, you’re trying to get a rough idea of what’s working, not fine tuning. And because your ad groups are focused tightly, you can make decisions on performance based on the overall ad group performance, instead of by keyword.
So if that ad group looks like it’s working, then starting spinning off other variants– add those keywords on other match types (assuming there is volume to justify it), come up another ad or two, maybe even bid a little higher. If it’s not working (based on my rule of spending twice the CPA amount), then you need to ruthlessly cut it. And if you’ve run a campaign for a few weeks and there are keywords that have no impressions or clicks— cut them. Don’t be afraid to prune. Same with cutting ads– look at both CTR and conversion rate. The search engines are deciding which ads to show based on what’s making THEM more money– which is based on eCPM. All else equal, Google’s optimization will choose ads that have the highest CTR, even though there is usually a corresponding trade-off in conversion rate. So when testing campaigns, choose the campaign setting to have ads rotate equally– don’t let Google choose. Finding the right balance between making Google money (to get more volume with high CTR) and maximizing your margin is hard. There is a clear trade-off between volume and price— as you have to bid up (or favor higher CTR ads) to get more clicks. Would you like to have 100 clicks a day that net you 20 cents of profit each or 500 clicks a day at 5 cents profit? I’m taking my first semester of Economics right now, and the teacher tells me that trade-off is elasticity– a measure of how much you need to pay proportionately compared to proportionately how much more volume you can get. Your profit is how many clicks you get times how much you net per click– it’s an inverse relationship, unless you are bidding on tail terms or perhaps certain branded traffic.
You’ll also want to look at your analytics data to get a sense of quality– don’t be just a PPC tool jockey. Check out your landing page bounce rate per keyword. You’ll find that some terms will have a 60%+ bounce rate and should therefore either be cut– or you have to change your landing page. A bounce rate is what percentage of folks bail on your landing page. Most affiliates choose to look only at conversion rate, but bounce rate is a great intermediate metric, since you get a lot more data earlier than having to wait for a conversion. We use our in-house analytics system, by the way, because we don’t want to give the engines our data. For many of our clients, however, we just install Google Analytics, since it’s easy, cheap (free), and has a beautiful UI. Anyway, the bounce rate is usually a good indicator of eventual conversion– after all, if they leave, they didn’t exactly have a chance to convert. So that will save you some money. If you’re sending people to your own site, you’ll also want to look at keywords that drive organic traffic. Provided you are not a one-page wonder and have a real site with information, then you’ll probably find a fair number of terms that people are coming in on— put those into your PPC campaign. And for terms that have worked well in PPC— start making pages on your site, so you can start ranking for them. I don’t look at things like KEI, LSI, or another TLA (three letter acronym). I rarely even use the Google Adwords API– but do in cases where there is enough volume to make it worth putting automated bid management in place. You do get dinged on using the API, for those who don’t know, so AdWords Editor is a more effective prototyping tool. Once you have something stable, then you can consider scaling it to the moon and using the API.
CREATIVES (ADS and LANDING PAGES)
Affiliates are notorious for copying each other– because if it’s working for someone else, then I ought to do it, too. Plus, it’s the lazy man’s approach. I admit that I do, it, too— since it’s a great starting point. But don’t just be a PPC tool jockey and think that this approach will bring you massive success. You gotta consider for a moment– if I’m doing what everyone else is doing, what kind of results can I realistically expect? It does frustrate me when other affiliates copy my ads— what am I going to do, tell them to stop or sue them? Guys, you know who you are. So come up with a clever twist for your ads and landing pages. That slight increase in conversion rate or CTR can allow you to make significant profits even if your PPC campaigns are no better than everyone elses. I highly recommend Tim Armstrong’s book on landing page optimization (insert amazon link– with affiliate code if you like). And promote related products on your landing pages— you already got them there, so might as well increase your chances of converting on something.
No comments:
Post a Comment