To bid or not to bid, that is not the question
A short time ago in a city not so far away (from Leeds anyway) i came up with what i felt was the Killer App for AdWords. This was a long time before i’d come across the long tail concept, chris anderson’s blog and my favourite Keyword Research tool of 2007, Hittail. Basically, the Killer App was a log file analyser that primarily focussed on the search terms that delivered traffic to your website. This tool would look at exactly what query delivered traffic to your website and tell you how unique it was.Why would i build a tool solely to do that? Simply becuase with the help of other pay-per-click bid management tools you can report on the ROI per keyword. Google, and now Yahoo and MSN, allow you to choose exactly what keywords and search term combinations you are bidding on and the ROI that each one, individually, is delivering.Looking for answers in the wrong placesWhat i frequently found, managing some very high spending campaigns, was that it was nigh on impossible to generate more traffic for a keyword without raising the bid price. Yet inevitably when a campaign was delivering a strong ROI, my client would always ask me if I could deliver more converting traffic to them, “Why don’t you just bid a bit higher?” In theory i could by raising the bid price, but in reality that strategy could be potentially disastrous, as not only would i be raising the ad position, to get more clicks, I was paying more for each click, so suddenly I might be in the position of kissing goodbye to that great ROI I’d helped deliver. Clients found my answer, that it was simply impossible to deliver better results simply by messing around with bids (after i’d found the sweet spot), difficult to stomach.Which left me with something to chew over. How could i squeeze a better ROI, whilst engaged in a bidding war with my competitors, without blowing my spend? I felt like i had to develop the art of zen non-competing.The PreambleThe idea of non-doing in Zen buddhism centers around a double denial of existence. Essentially it is the logic, “life is empty and meaningless, and it’s empty and meaningless that it is empty and meaningless”. Basically the logic posits that the statement itself is neither true nor untrue becuase the statement is a component part of reality that only holds a mirror up to what is actually going on. Put another way, the statement itself is only a reflection of what is going on, a description, not actually what is happening.Pure PPC strategies are just speculative at bestBelieve it or not, Pay-per-click bidding is exactly like that. Pure speculation. Most SEMs will push the idea of never before seen levels of campaign control and targeting, yet they are wilfully overlooking that fact that they are overseeing what is essentially a blind auction. When you choose to bid on the term, say ‘jobs in charities’ you can select that to be a broad, a phrase or an exact match, and your bid management engine will deliver an ROI value for each permutation. It’s great becuase it gives the feeling of confidence (especially when you see your exact match is delivering a great ROI), but it’s not necessarily true that the term jobs in charities is the best converter. This is especially clear when you notice your broad match is consistently delivering strong ROI too. A more savvy PPC manager should find cause for concern regarding the long term health of their campaign, simply becuase they cannot say with the same level of certainty what terms are driving conversion. And therein lies the opportunity.Finding answers in the right placesTypically, although this does vary from market to market, you’ll find that your exact matched term delivers the best ROI but fairly quickly you’ll start to see that drain away as your competitors start competing aggresively and driving your bid prices up. A typical response by most agencies i’ve worked with is to add more keywords and variations to fill out the bottom end of the market with search terms that have lower cost-per-clicks (CPCs). Frequently this strategy works, but it does not solve the problem of what to do about rapidly diminishing ROI from your super-converting term. Furthermore, in most cases, these ads deploy ineffectively because the click-thru rate on your broad match, out performs your new untested and unserved keywords. So you we’re back to square one, becuase the broad matched ad was served in their place and we have no new data.So how do you solve this problem?Firstly, optimise your clickthrough rate (CTR) rather than your bidding strategy by blocking certain terms.Find all the irrelevant matches that are coming in and delete, delete, delete! Set up negative matches on your campaigns so that your ads don’t serve for any irrelevant phrase or broad matches. This should reduce the number of impressions you get on irrelevant terms, pushing your CTR up and furthermore, becuase you are not getting clicks from terms that are unlikely to convert (hey, you know what useful content is and is not on your site), you are optimising your ROI.Secondly, optimise your CTR rather than your bidding strategy by adding new terms.Discover what relevant variations are still coming in. Log all of the long tail combinations and variations that are coming and see what juicy stuff you can learn about your market and where you are failing to serve them effectively. You will honestly be amazed at the granularity of search terms clicking through to your website in the hope of finding something relevant. For instance, in the broad matched example of ‘jobs in charities’, I was amazed to find how much volume there was to the term ‘jobs working in animal charities’ and variations to that effect. Yet, if you think about it for a second, with a global database and the ability to search it at 0.11 seconds, it stands to reason that the more inquisitive user knows exactly what they want and is going to set about finding it!Thirdly, optimise your CTR rather than your bidding strategy by restructuring your campaign.Re-structure your entire campaign to accomodate new learnings, in such a way that you can learn from the new initiatives you are taking. PPC data should be seen as source of response data about your audience. Aside from having some assurances as to what search terms deliver business, clickthru rate metrics give you some pretty good assurances as to what your audience really reponds to. A lot of AdWords accounts that i’ve taken under my wing to optimise basically fail to deliver more becuase of the way they are structured. High traffic terms are under the same budget control as low traffic terms and in many many cases one generic ad is serving a multitude of different search requests. This is really problematic because you are unable to get a clear overview of what keywords and ads your audience responds to. Restructuring your campaign really allows you to know at a glance how relevant your ads and website your website is in general to the market your targeting. People find themselves in different terms. for instance I principally think of myself as an SEO and PPC expert, rather than an SEM expert. Yeah, i may be quibbling over definitions, but frankly if I am, then your audience probably are too.Even if you don’t bother with steps 1 and 2, a campaign restructuring will deliver more impressions, have a much less volatile cost-per-click (CPC) and higher CTRs. A solid account structure will give you far more control over your budget on a long term basis to engage in or opt out of particular bidding wars. On a daily basis you can switch on and switch off campaigns and adgroups whilst your competitors bid aggressively, until they burn out and you can slide back in.Fourthly, optimise your CTR rather than your bidding strategy by writing more specific ads.Serve your audience better by serving better ads that meet their specific requirements. Until you re-structure your campaign to accomodate the new keywords, you’re going to be under the impression that your broad matched campaign delivers you a fairly standard number of impressions at a fairly standard 3% CTR at whatever CPC. When you restructure your campaign, and serve a relevant ad (and please don’t forget a relavant destination webpage) to that query, you’ll discover that in fact that particular term, say ‘jobs in animal charities’, drives over 500 impressions a month and delivers a 25% clickthru rate when targeted with relevant ad text and landing page. Sit back and watch your ad sail above your competition without you ever adjusting your bid price.Fifthly (is that a word?), optimise your CTR rather than your bidding strategy by continuously improving your account.Repeat the whole process on your new campaigns and adgroups. Each new campaign and adgroup you target will deliver more opportunities to discover how you can compete in the marketplace with your current budget and where to best position your spend.Search terms ARE NOT markets in their own right, only reflections and sources of insightsSo where is this killer APP now? Well the long and the short was i was too young and inexperienced in the SEM business at that time to really win the funding i needed to develop it beyond it’s neo-natal stages. In the end the business died (archived site here) because i still needed to support myself - but hey, it’s totally OK to fail at setting up your first company! (Seriously, the UK needs to realise this and learn from our bigger brothers across the pond). Anyway, the most important thing is, as the dalai lama is alleged to have said, “when you lose, don’t lose the lesson”. And I can safely say that i didn’t. Instead i put a similar tool, Hittail, into action with some splendid results. So when you think you are losing the bidding wars (and the ROI), remember not to lose the lesson - ask yourselves or your agency how far you are going to improve clickthrough rates.And btw, Hittail have just upgraded their service. Isn’t that nice!