London SMX Advanced: Keyword Research – Beyond The Ordinary
This session looks at innovative approaches to keyword research that help you find those really significant keywords your customers are searching for, uncover your competitors’ keywords, mine the long tail for those elusive but truly valuable obscure search terms, and expand your “virtual shelf space” in surprising and non-obvious ways.
Moderator: Kristjan Mar Hauksson (KH), Director of Internet Marketing, Nordic eMarketing
Q&A Moderator: Merry Morud (MM), Search Marketing Account Manager, aimClear
- Andy Atkins-Krueger (AA), Managing Director, WebCertain
- Christine Churchill (CC), President, KeyRelevance.com
- Sam Crocker (SC), Lead SEO, Distilled Ltd.
- Barry Lloyd (BL), Search Marketing Evangelist, Unica Corporation
(The below are Tim’s notes from the Keyword Research session at SMX Advanced. They have been posted during the session and will be tidied up later – more SMX coverage here)
Data Mining Technique related to the Editorial Calendar – a basic but oft-overlooked technique.
- Editorial Calendar shows major planned stories for up coming issues of magazine or newspaper [additional idea – look up editorial calendars of those in industry]
- Morph it for keyword research by identifying scheduled events (conferences etc), seasons, trends etc with relevance to your industry. Can then plan your blogging/twitter updates, PPC, YouTube videos etc
- Example: RenFest (Renaaissance Festival) where people dress up as kings and queens and speak in bad english accents. Lot of search volume in run up to event. Can also use Google insights – regional info very useful in States.
- Use excel to draw up appropriate lists of events etc
- Many suggested tools including Google Tools, usual paid ones, Twitter trending topics, YahooBuzz
- How can you capitalise on the ashcloud?
- Consider making seasonal content ever green
- Have content waiting in the wings
- Think well ahead – also helps you be much better organised and unified.
Using broad and phrase match for your keyword research
- Analytics data is crucial info often overlooked.
- Finding keywords that convert is vital. Many phrases get clicks, far fewer make money
- Move converting broad match keywords to exact match as cheaper
- Advocating use of paid search as a keyword research technique. Likely to find phrases that other keyword research won’t kick up.
- Most people don’t get down to actual search phrases in Google Analytics as default shows bid phrase
- Ultimately you should look to increase number of exact match phrases as an AdWords account matures.
Becoming a Keyword esearch Ninja
- Standard tools all OK, but other tools described below increase efficiency, none are designed for keyword research
- Mozenda – a tool designed for crawling and scrapping. Google XML API to get Google’s suggested related terms out efficiently. Use Mozenda to take existing 1000 phrases and generate more. Navigate to Google suggest API, import an excel file with existing phrases. Can also use Mozenda with Delicious to pull out tags associated with a keyword. Also use Hulu (example was with a TV series)
- Most popular TV shows – didnt follow the point
- Xenu Link Sleuth – crawls sites, strip keywords from urls
- Microsoft IIS (Xenu on crack)
- Mechanical Turks – outsource data
- 80 Legs – similar to Mozenda but more user friendly
Keywords can not be translated! Andy is very keen on Casseroles! In French they are pots for cooking, in english they are stews.
- Concerned by our reliance on keyword tool
- Google translate (can use the detect feature to determine how Google views a word. For example it detects casseroles as being a french word)
- Keyword tool is giving different numbers depending what language you put in. So don’t believe everything.
- The longest search tails are actually in Dutch in German, not English. Spain has a very short tail.
- How does Google compute Google Suggest Suggestions? SC: Something to do with number of searches for phrase. AA: Manipulated to have relevance to user. Could be related to IP but it does produce different results depending on what domain using. SC: Search history seems to be a factor
- Competitiveness of keyword tips? AA: Not a fan of KEI (Keyword Effectiveness Index), a much better approach is to do an intitle search and look at the volumes on that. SEOmoz has some metrics but as with any metrics need to take care. KH: Recommends Trellian
- Use own web analytics as a tool? CC: tendency not to go deep enough, filter out brand, look for two to three word phrases that incorporate into longer phrases. AA: Look for phrases in our Analytics that you don’t have first page listings for. Likely to be gold dust. AA: Bounce rate under utilised metric.
- Other tips: CC: Yahoo suggest is slightly different and may give other ideas.
- How to judge long tail phrases with one conversion – good conversion or not? AA: Need to look at patterns.
- Qualifying keywords to decide if you need additional content? CC: Is it relevant? What is user intent? Need to make a call, somewhat subjective. AA: Run a keyword map for your site. With new keywords can then look if you have the content already and can just tweak.