Five Steps to Optimization Five Steps to Optimization Beyond Best - - PowerPoint PPT Presentation
Five Steps to Optimization Five Steps to Optimization Beyond Best Practices Beyond Best Practices Beyond Best Practices Beyond Best Practices Sokolove Law proprietary and confidential 5 Steps to Optimization 5 Steps to Optimization
Sokolove Law proprietary and confidential
g
Sokolove Law proprietary and confidential
2
CPCs ell o er $200
Sokolove Law proprietary and confidential
3
Sokolove Law proprietary and confidential
4
Sokolove Law proprietary and confidential
5
Do you or a loved one have a mesothelioma diagnosis? Find out if you’re legally entitled to compensation you re legally entitled to compensation.
Sokolove Law proprietary and confidential
6
Result: $109 for the click, no lead.
Sokolove Law proprietary and confidential
7
Sokolove Law proprietary and confidential
8
St 2 Wh th i j d t d t b t ? (3
Sokolove Law proprietary and confidential
9
Sokolove Law proprietary and confidential
10
Sokolove Law proprietary and confidential
11
Sokolove Law proprietary and confidential
12
Sokolove Law proprietary and confidential
13
Cara deBeer
SEM Manager
Slide 1: BEYOND BEST PRACTICES for PPC Good afternoon. I’m here to talk about 5 ways to optimize for paid search. I’m sure I’m not the most experienced marketer in the room, nor do I have access to the paid search algorithm. But as Isaac Newton said, If I have seen a little further, it is by standing on the shoulders of Giants." – and I’ve been lucky enough to work with some very talented people who were generous about sharing their knowledge with me. I’ve been doing PPC for the past six years and I’ve run into some really weird problems in that time, some of which I was able to solve creatively and some of which blew up in my face. Standard best practices can only take you so far and so today, I’m going to go beyond best practices. SLIDE 2 – 5 STEPS Also with me today is Jeff Eckman, from BGC (big giant conversions). We’ve been working with BGC for about two years now, partnering on LPO (landing page optimization), and it’s been a very important part of our success strategy. While I will be sharing some concrete tactics, the main thrust of this presentation is to inspire you to find your own ways to bend the rules of best practices. I’m focusing on Google Adwords today because I don’t have a ton of time up here and let’s face it, Google provides the lion’s share of traffic to most advertisers.
When I interviewed for my current job – which is essentially lead gen for personal injury law firms ‐ my future boss asked me if I was familiar with mesothelioma. I knew it was a rare and deadly cancer caused by asbestos, but what I didn’t know is that it’s reputedly the most competitive vertical in paid search. SLIDE 3 – OPTIMIZATION IN COMPETITVE VERTICAL ‐ Ever since Google removed the cap on bids, I’ve been paying $250 per click for some keywords ‐ Quality scores of 3 are GOOD. ‐ Low volume, high margin industry means data scarcity can be a stumbling block for analysis Best practices are a great place to start, but that’s all they are: a starting point. I read a lot of industry blogs, and no one ever talks about what the heck you’re supposed to do when you’ve inherited an disorganized account with two years of spotty history, a million keywords, and (in which) the only quality scores higher than 4 are on the brand keywords! I also need to spend several million dollars efficiently, with a lead cycle time that usually lasts two weeks but can take months or even a year. That’s right – that click from July 2010? It still matters. Optimization #1 – QUALITY SCORE My quality score problem is a little more extreme than most industries, but I’m sure all of you have keywords which you know could work better for you, but which you just can’t budge the quality score on, and so you’re priced out of the market on that keyword, and you’re losing
SLIDE 3 – QUALITY SCORES In fact, I haven’t been able to find any articles that address my situation – they all tell you to get rid of keywords with quality scores that are 3 or below. Yikes – that would be most of my account! I’ve got instances of “mesothelioma” with a QS of 1. (The slap.) And these keywords are converting (when the ad shows). Bidding on keywords with a terrible quality score is as expensive as you’d expect – but the ROI in this high margin, low volume industry means that if I get a lead that converts, a $250 click is chicken feed. I’m lucky enough to have a dedicated Google team for my accounts and my reps tell me that these low quality scores are pretty standard for the industry. (In fact, I leaned on them to provide me with some anonymized competitive information, but the results were disappointing: we were doing as well as or better than our competitors in most areas. It’s not that I was hoping to find out that I’ve been doing a rotten job, but it would have been nice to get some actionable insights.)
So I’ve got a pretty vested interest in figuring out more about how quality score works in this vertical, because even though quality scores of 2 might be normal, I still don’t want to pay that much per click – I’d rather improve the QS and use the “left‐over” money for testing. SLIDE 4 ‐ FRUSTRATION I think it’s pretty widely known that there are actually several quality scores: at the keyword, ad group, campaign and account level. But: did you know that QS is also attached to the domain? Let’s say you start a new job. It’s pretty common to arrive at a new job and be handed some inherited accounts, with mediocre or flat‐out bad performance history, which have been badly managed for a while. (That’s why they hired you, right? To fix things.) If you decide to scrap that account and start fresh, you’re also going to need a new domain, or those crappy quality scores will follow you across the adwords network. And then your new boss will probably be annoyed that you put all the effort into creating a new account from scratch only to have it not make a difference. Trust me: there’s nothing more frustrating than having creating adgroups of less than ten keywords, all on a very similar theme, which are paired with painstakingly written ad copy which incorporates most or all of the keywords in the adgroup, which send traffic to carefully crafted landing pages that not only include the keywords, but also mirror the ad copy – only to find, at the end of the day, that you STILL have a quality score of 2 with only a mouseover “poor relevance” to guide you. (Screenshot). It’s even more frustrating is when you’re getting a 2% CTR on that keyword. 2% isn’t high in all industries, but it’s high for mesothelioma. It’s enough to make you tear your hair out trying to figure out what else you can possibly do to improve your account and bring down the CPC. OPTIMIZATION #2 –ACCOUNT STRUCTURE AND ORGANIZATION So here’s what you do. You take best practices, and you square them. Google recommends tightly grouped keywords‐copy‐landing page? Fine –
don’t recommend it unless you’ve got a fairly small number of head keywords driving the majority of your traffic. But it *will* get you crazy high CTRs and that will improve your quality score and drive down your CPC.
Another commonly recommended best practice is to build out the long tail – the theory being that if it works, great, and if not, it’s not going to do your account much damage. As a matter of fact, the long tail isn’t harmless if it doesn’t perform. Yes, it makes 100% sense to go for the long tail if you’ve got a pretty vast universe of keywords – say you’re in retail, or travel. But if you’re in a competitive lead gen‐service area like finance? There just isn’t much of a long tail on highly relevant, transactionally oriented financial service keywords. That’s not to say that you can’t build one out, with a million iterations – you know the drill ‐ include synonyms, cross‐combine all your variations with geographical modifiers, use every possible match type and track them all to see what works best. SLIDE 6 – Suspended keywords with low search volume The problem is that if you’re already in an area that’s extremely competitive, having a whole lot of long tail keywords that languish at the bottom of your account in a “suspended” state most of the time, occasionally waking up enough to garner impressions but not active long enough to get clicks – well, a few impressions over a month on millions of keywords which never get any clicks isn’t going to do much for your
to navigate. It seems almost sacrilegious to say this, but sometimes you can do more with a tightly controlled head of 30 keywords that you watch like a hawk than you can with five million low search volume keywords and a bid management system to spread the workload. I’m not advocating putting all your eggs in one basket and only going after a few head keywords, but I’m pointing out that you need to be proactive about managing the long tail – it’s not a good idea to set it and forget it. I mentioned match types as a keyword interator, and another best practice seems to be bidding on all match types (one match type per adgroup), and then tracking to see which performs best. But depending on the field, it may not make sense to run on all the match types. You might not need to do the work of setting that up and analyzing the results. For example: not that many people one can spell mesothelioma, so phrase match doesn’t really make a lot of sense for me: I’ll catch the good spellers with an exact match, and I need broad for the others so that I can pick up misspellings and related keywords like “asbestos lawyer”. (Asbestos is another word that people have trouble spelling, so if I were going to run on asbestos, I would need to run on broad. In fact, running “asbestos” on broad match turned out to be an excellent way to spend money quickly, with little to show for it. (I might as well have gathered the money in a pile and set fire to it.) )
keyword which got the credit for the click, sometimes it’s such a wild leap that it leaves me shaking my head.
SLIDE 7 Recently I had the pleasure of paying $109 for a click on “mesothelioma symptoms” when the raw search query was “tylenol for fever.” Session‐ based broad match was the culprit. Presumably, this person had been doing asbestos or cancer‐related searches in this session before the Tylenol search … but that doesn’t matter to me. Tylenol for fever is never going to be a good lead gen keyword for a campaign which focuses on finding mesothelioma patients who want to pursue personal injury lawsuits against the manufacturers of asbestos. So how did I address this? Modified broad match, which only includes close variants – i.e. misspellings. I created an identical campaign but put those keywords on modified broad and set it to run side by side with the original broad match. I couldn’t just switch over 100%, because I need to balance lead quality and cost effectiveness with volume, and modified broad will cut out a lot of volume. I know that using modified broad match in conjunction with search query reports isn’t exactly revolutionary, but I mentioned this example because I was using search query reports in a slightly unusual way– most of the time search query reports are suggested as a means to build out your keyword set or to add negative keywords. But I already know my keyword set, because it’s pretty small, and even with an 800:1 ratio of negative keywords to positives in every campaign (which is the ratio that I actually have), I’m still never going to be able to cover all the variants that broad match could possibly suggest. Instead I compared the search query report to the keyword report to find out more about what was happening under the hood of the algorithm, and that helped me to come up with a solution which was more cost‐effective while not sacrificing too much volume. 3 – LANDING PAGE OPTIMIZATION The same principles apply to landing page optimization. Mostly the best practices for this stick to “write relevant headlines” “customize as much as you can” and “conversion form on the top right”. Okay, awesome, I do all those things too. But that’s not enough. And I’m not a designer nor am I a useability expert, so I make up for that by testing and breaking the rules. SLIDE 8 SEOMoz and the Conversion Rate Experts did a terrific case study where they found that long landing pages actually improved conversion rates. I thought that was a great idea. We run 30 second TV commercials – why shouldn’t I test a landing page that takes 30 seconds to read? Lead quality matters so much in my industry, so it makes sense that long landers could help visitors to gain a sense of trust in our brand and the service we offer, pre‐qualifying themselves at the same time. But a long landing page is completely against best practices.
SLIDES 9, 10 & 11 – MULTI STEP LANDERS The idea of making visitors go through multiple steps on a landing page is another idea that flies in the face of best practices ‐ but when we tested a landing page experience which funneled visitors through several pages, answering questions designed to help the users self‐select, we had much higher form fill rates (for the users who made it all the way through) than we did on a single page lander. OPTIMIZATION # 4 – TESTING SLIDE 12 ‐ FAIL I’ve talked about testing a lot. This is another of everyone’s favorite best practices. Testing is great. The part no one tells you is that 9 out of 10
very good to carefully evaluate the business opportunity of a test, invest resources in developing, monitoring and reporting on the results and then … watch it bomb. And then have that happen ten times in a row. (The odds are still better than Las Vegas.) So I end up giving myself a lot
motivational talks can only take you so far. I address this by testing constantly. Then as a test fails (or …not), I’m already developing the next one and I’m less emotionally invested in the
addressed very often … it’s often pretty easy and low‐energy to test things like ad copy, or match types, or landing page messaging. But if you’re testing constantly, it’s easy to set tests and then forget to keep track of them. Can this be easily avoided with meticulous reporting? Absolutely. Do I spend 50% of my time in meetings, rather than working on optimization or reporting or tracking? You bet. So when I set up a test, I also set up my reporting formats beforehand, which makes goals of the test clear. Then all I need to do is plug in the actual numbers, and my analysis is three quarters of the way to finished. Test some off‐center ideas. Try not putting a call to action in your ad copy. If it’s a highly transactional keyword, such as “cost of blue widgets” you might not be served well by putting “buy widgets now” in your ad copy – users might respond better to an ad which highlights your terrific selection of blue widgets at bargain basement prices. Of course, you can take this too far. I once had a Google team write some ads for me which were designed to run as text display ads and therefore, they needed to be attention‐grabbing. The ad they came up with was
Online Backup Tastes like chicken! … which was attention getting, yes, and very playful. It was also wrong for the brand, which had its demographic of older white men, highly educated, and over 35. These were people who took themselves seriously, and they wanted their online backup company to take their data backup seriously, too. You know what, though? I went ahead and tested it anyway. Just because *I* thought it was a terrible idea doesn’t mean it actually was. What do I know? These ads were written by Google display specialists! Experts. From Google. (Also, these were display network ads, so the risk of one of the senior execs coming across a completely absurd ad, and firing me for such a stupid idea, while vanity searching on a Sunday morning were pretty low.)
SLIDE 13 OPTIMIZATION # 5 COMMUNICATION How to communicate PPC to others at your company, if you’re in house. The high visibility of PPC means I spend as much time explaining to execs what exactly it is that I’m doing in those spreadsheets as I spend on strategy and execution. This less of a problem at an agency because if an agency has a PPC engagement, it’s because they’ve sold the concept to the client, which means the client has a basic understanding of either what goes into PPC or what paid search can do for the company from an ROI perspective. No one just hands an agency a bunch of money and says, “go blow this on paid search, we don’t care if your fees plus the upfront costs put us in the red.” But even so, working at an agency, you still need to get cooperation from others in the agency. You need internal buy‐in on your testing, or else you’ll find that your access to test development resources – everything from design to back end tracking to hosting – is blocked, and your testing will be limited to what you can test through the engines’ test options. So do your best to cultivate relationships within your organization, because that’s crucial to the success
positive feelings. )