The hall is still too hot, but it’s been another good day of presentations with a slightly slower overall pace. Here’s the day one Search Love review if you missed it.
Excuses: Please excuse the lack of checking these posts (they’re posted on the day) and my lack of en-dashes today and yesterday; the Chromebook can’t do them (at least easily). It makes me sad. Also apologies for the terrible picture.
Building a brand with online video – @philnottingham (Distilled)
Video success has sometimes been recorded by impressions, but impressions aren’t really making an ‘impression’ on anybody.
A good video brand needs style (recognisable), content (that’s different), distribution (to get to the right people) and measurement (to see how well it does). Phil’s big message is to play around a bit with video! Always consider your:
- Style: performers that reflect the personality of the business (don’t use stock!). perspective (first person, second person etc), palate (colours, tone and restrictions), personality (use the language of your audience – know which of ‘the big five personality traits’ your user is).
- Content: What works on TV might not work on YouTube. You need to tell a great story – but how? You need customer insight (use Google Trends, Tubular for ‘videos users also watched’), competitor insight (Unruly Analytics is like SearchMetrics for YouTube), and ‘product truth’ (what people think about you: SurveyMonkey, Google Consumer Surveys).
- Distribution: You need to be on YouTube – simple as that. It also needs to be shareable, but things don’t ‘just go viral’. Everything needs a massive push at the start, but paid distribution is not a shortcut (although it can help to amplify). You can also do some competitor research to see who shared other videos and outreach to them. How do you know if it’s going to be shared? Well you can pay shed loads for tools or just read up on sharing psychology. You could also use UnrulyActivate to push videos out to bloggers etc. If you’re running pre-roll YouTube ads, grab users with surprises, subvert expectations and use your knowledge of your audience to target placements. Phil says if you don’t have a YouTube strategy then you don’t have a full SEO strategy.
- Measurement: Phil says report on active metrics like share rate (views divided by shares) or ‘engaged views’. Also measure assisted conversions, which tracks if somebody eventually ended up purchasing or signing up through other channels e.g. remarketing. Remember that last-touch attribution doesn’t work for branding and branding videos; you’ll need to use another model and at least consider first touch. Phil’s also not too keen on measuring how much of a video has been viewed, as the ‘impression’ can be made at any point.
Rich snippets were discussed in the Q&A: they’ve been largely removed now, so YouTube has to be considered more strategically.
Grammar nazi review: Inconsistent capitalisation and a lack of full stops let Phil down a little. Otherwise a typically energetic and captivating performance.
Mobile analytics – @justincutroni (Google)
Data is the cornerstone of our marketing efforts: we’re OK at tracking the mobile web, but Justin asks whether we will move towards HTML5 or apps? How do we measure app use? What about people moving between the mobile web and apps? Mobile is tailoring the experience for the information about the user at that point in time (so using location, camera etc.).
Unfortunately app data is split between the Apple App Store and Google Play.
As with all marketing, you need to have a plan! Set objectives: generate app installs, encourage in-app purchases, and keep users engaged. Then set KPIs against them, and segment the data you’ve collected.
Segmentation is key, as you need to identify groups and move them towards defined goals. App stores also collect some demographic information, so don’t forget to use that. In terms of improving the returns from existing users, you can target users using on-phone messaging systems and cross-app advertising.
SDKs (software development kits) are what’s used to gather app data generally, but you can actually use Tag Manager for apps. That’s pretty great; we’re fans of Tag Manager at Attacat. On mobile users are defined by a user ID, and this is far more accurate and reliable compared to cookies. This is generally anonymised and passed to the advertiser.
In apps you need to use events to track interactions, and it can be a lot of work. Don’t let the developer guide this process as you need to source meaningful data. It’s quite hard to track app installations, as you drive people to a store that you’ve very little control over, but Google Analytics and Play Store developer accounts tie up quite nicely.
If you want to learn more about app analytics, try Google’s upcoming app analytics training academy!
Grammar nazi review: Next to no text, so no issues. Justin was an excellent presenter and I’m a bit of a fan.
Putting the X into content – @mollyflatt (1000 Heads)
A lighter diversion in this conference, but a little light on value for my money. The overall message was that consumers are becoming content addicts; we’re pretty fed up with it but we’re hooked. Marketers are just ‘content pushers’, feeding an addiction with crap content – content without content.
What can we do about it? Start thinking ‘context marketing’ and targeting influencers, as content is less important than the social action it performs. Content should generate a response and get users to ‘feel’ something, and to do this you need to understand the needs of your audience at a particular moment in time and be useful. Focus on generating emotion.
Good social and content marketing is understanding that the content produced by users is much more important than what you produce to get the conversation going. Try reading the book ‘Contagious’. Although it’s probably a risk reading it on the train with the current Ebola scare…
Grammar nazi review: Said ‘recreate’ with a short vowel once, which was quite weird. Consistent use of full stops, but there were a couple of typos and some inconsistent capitalisation. A chirpy presentation that seemed to me, ironically, light on actionable content.
Turbocharging your WordPress – @jonoalderson (LinkDex)
A “chaotic list of tips”, interrupted by a fire alarm and building clear-out, tied together by the theme that we should care about technical SEO. Technical SEO is not on-page optimisation – it’s server configuration, scripting languages, site speed and more. I’d suggest reading the slides if you’re interested in this talk.
Why should we care? Well page load speed is a ranking factor and can massively impact revenue; schema, canonical and Href lang markup can have massive impacts, 301s and site errors can ruin your site and search presence. So, sort out your page speed using any number of the tools available. Unfortunately, developers hate understanding it and fixing it…
Jono identified seven distinct technical SEO areas that are being, somewhat terrifyingly, left behind, followed by a list of WP plugins and third-party services that cover most of the things that need doing. What about Yoast? It’s great, but it can be dangerous: you can miss errors and stop you thinking about the ins-and-outs of technical SEO.
- Hosting: Go for a VPS, or consider supported services. You can also play around in the CPanel and WHM interface.
- Custom error pages, security holes, unnecessary templates and author pages.
- Theme performance: use W3 Total Cache – it’s great.
- Redirection: use the ‘Redirection’ plugin, as it shows you how often redirections are requested.
- Security: use iThemes Security.
- Server location: it matters. Use a CDN like CloudFlare. Also by putting things on subdomains you can load more things in parallel, and this can be set up in CloudFlare. Also you can use established CDNs to serve things like jQuery.
- Minification: use WordPress Minify.
- Secure server: It’s apparently really easy to migrate to SSL. Change to protocol-relative (e.g. //example.com/file/). You can also do SSL through CloudFlare! You can also do a cPanel database query to find old links that need updating.
- Media: pay attention to the details, and don’t upload huge files! Remember that Photoshop is bad at exporting well-compressed jpegs. Try specifying different images for different browser widths to reduce load.
- First-bite: the first connection is key.
Phew, there are too many tips and my fingers hurt…just look at the slides (I’ll insert them here when they’re available). Also did you know tomorrow is both cat day and internet day? Check out daysoftheyear.com, Jono’s site.
Grammar nazi review: Noticed one ampersand where he should have just put ‘and’ and a typo, but maybe one of the best over the two days and I can’t nitpick given he did a fantastic job after being interrupted.
What black boxes can tell us about the future of search – @kelvinnewman (Rough Agenda)
Kelvin thinks he’s getting worse at understanding what Google wants.
He talked about evolving circuits to solve problems, and how they’ve evolved in ways that we can’t reverse engineer or totally explain, and a flash-crash in the US stock market. The lesson is that incredibly complex algorithms can react in strange ways in the real world and in combination with other algorithms.
And this takes us to search: Google’s algorithm is incredibly complex and works within a complex and chaotic system, and Kelvin suggests that Google is actually a machine-learning company. Google therefore cannot know how an algorithm update is going to react when let loose in the wild, wild world, and neither can we.
Just because SEO studies of the algorithm are interesting, it doesn’t mean they’re useful. It seemed to me his message for SEOs is: don’t try to reverse engineer the algorithm, just “do what you do” as a good search marketer.
There are a few books for your reading list: The Signal and the Noise, Flash Boys and (oddly) Jurassic Park.
Grammar nazi review: Punctuation but seemingly random initial capitals? Typos? Horrid use of ampersands? Incorrect punctuation inside a bracket? Aarghhh. I love a good semi-colon but I’m afraid it was used incorrectly multiple times. Typographically it was …’eclectic’. I’d say stick to the all caps – sorry Kelvin. Overall, an engaging theory-based presentation.
Girlguiding social strategies – Jo Kerr (@Girlguiding)
A bit of a change of tack from Girlguiding, which was interesting for me as my partner @AlexEdits volunteers for them. They’ve developed a new digital strategy to involve the girls, enable and champion the Girlguiding story, and Jo talked us through some of the elements of including persona development and understanding the attitudes and opinions of their target audiences. Actually speaking to their audience was emphasised as a big part of the organisation’s success.
Jo also stressed the importance of social media for transmitting messages and connecting emotionally with users.
Grammar nazi review: No notable issues as it was mostly screengrabs. Jo was a good speaker but clear, focused insights were somewhat lacking for what is a knowledgeable audience.
Content: embrace your inner geek – @mattbeswick (Hidden Pixel)
Matt is a geek, but he thinks he’s a fake geek. He’s not a coder or a developer; he just hacks things together. He’s still not averse to including a fair amount of code in his presentation though!
He wants us to remember that the specific tools might not matter, but thinking about what you can do to automate your processes, get cool stuff and move towards becoming a brand.
Unique data is great for outreach! Surveys help to gather together interesting data stories, with story being the key word. Matt recommends using Google Consumer Surveys, although OnePoll is more respected by mainstream media. Get your data and pull out what are the interesting findings that creatives can then work their magic on.
Matt talked about scraping and APIs – both are very useful to know about and it’s well worth knowing how to form API requests.
Getting this information for one data point is OK, but getting it for loads is where is gets interesting. You can pull out Twitter follower data and see the sites of influential users in an instant. You can also work out the sites that they’re sharing or reading. It’s something that Richard Baxter talked about at MozCon last year so it’s something I’m very familiar with, but it’s a great tip to remember.
With APIs like SEMRush you can see what keywords your competitors are bidding on, and using URLprofiler you can even automatically take screengrabs.
But let’s talk about links. Use MajesticSEO API, grab the links, filter by those linking to competitors and boom you’ve got a lot of solid link prospects. Perhaps use TextWise and spider the web for relevant sites about particular topics.
Grammar nazi review: Oh my God I hate initial caps on every word. Please don’t do it. Please don’t also sometimes do it and sometimes not, as that’s even worse. Some appropriate hyphenation was nice to see as a saving grace. A good speaker and an interesting topic.
Running promotional campaigns – @iainhaywood
A competition requires skill, a giveaway requires no skill to win. This is where we start this run-through of competitions.
There are three golden rules to remember:
- There is more than one type of entrant: one is primarily interested in winning, while another is primarily interested in your brand and vertical. The latter are the dudes you want.
- It’s not a magic bullet: it won’t drastically impact your revenue.
- Incentivisation totally changes the nature of intent: they’re doing something because there’s a prize at the end of it.
What are competitions good for? Email and other data opt-ins, social media, coverage, basket value increase, SEO etc. By getting value in more than one way you have a better chance of getting a decent return on investment.
Treat the running of competitions like a PR exercise, not a linkbuilding one, and go for quality. If you’re partnering with third parties make sure you get all the opt-ins (or social shares) as well as the hosting organisation. You (hopefully) also get a link. For a first-party competition, include incentives for sharing as this can help to promote the competition without your ongoing input.
Convey trust by having a ‘technical’ competition entry: something that clearly tracks the entry so the user knows their entry is properly considered. Iain suggested tools like RaffleCopter. If you’re running a Twitter competition, use something like TwitterDraw. If you’re considering running a voting competition, then don’t. It’s apparently too prone to being manipulated and can end badly.
OK so you want entry volume? Look for competition communities, forums, competition bloggers and influencers. Don’t forget to promote your competition with display advertising if you’ve the budget.
Unsurprisingly the prize makes a huge difference: if it’s ‘transformative-level’ then you’re on to a winner. Then get some famous judges of some. Craft your competition around “What would the Daily Mail write about?” and that will give the PRs/journos something inane to cling to. Iain says get yourself a media partner: basically a big name you can hang off. Just ask them to be a sponsor, and when they say “No” say “How about being our official media partner instead?”. Easy.
Watch out to ensure you’re compliant in whatever country you’re running a competition. UK is quite soft apparently, but specify what countries/areas can enter and check the data protection act (CAP code section eight). Be careful of automated entry services and other fraud methods, and try to cover for it in your T&Cs.
There were many tips suggested, and one SEO one to sooth the attended masses was to host the competition closest to the key page (e.g. homepage) you’re looking to optimise for. Finally, for prizes make the equivalent cash value clear and don’t offer tickets!
Grammar nazi review: Full stops were in short supply but the (seemingly typical) inconsistent use of capitalisation wasn’t. Double spaces after full stops drive me insane – do you mistakenly think you’re using a typewriter? Interesting trouser/shirt combo but a nice presentation.
Cracking the SEO code for 2015 – @randfish (Moz)
- Searchers and search engines are demanding more from us / websites: 47% of consumers expect load speed <2 seconds, users ‘bounce’ away from poor content, and there are now human quality raters. What used to work probably won’t work in the future.
- We’re moving from keyword matching to topic association: Google has improved its understanding, but has our content improved alongside?
- Domain-level keyword connections are more important: it’s becoming about supporting the brand rather targeting the keyword. The issue is that if you’re not a brand, Google is likely to screw you at some point for your SEO tactics.
- Some referral/keyword data is still available: but it’s only useful for high-traffic sites (and brands?) who are spending shed-loads of money.
Keyword research
Instead of using keyword research tools, search the web (news, images, video, buzzsumo), collect together concepts and topics, and aggregate into a new keyword list. Instead of grouping keywords by ability to rank, group by overlap in searcher intent. Instead of creating pages by keyword, produce pages for search intents. Instead of worrying about keyword cannibalisation, make more content on those themes to build domain relevance.
Google’s keyword tool has search volume data but will only show it for the high commercial intent keywords; it won’t show you volume and suggestions unless you enter the exact keywords you want to see.
You need to build a site’s association with a topic or range of topics.
Linkbuilding is now link earning
Remember it’s about building a relationship in order to get a link eventually. Social, email, comments: all can lead to the building of a relationship where a link follows. Consider that the ‘top-rated’ guys will get approaches all the time, so consider approaching the nice, fat middle ground and they’re likely to be a lot more malleable to your intentions. Try GoConspire if you’ve got a good range of contacts, as it’ll tell you how you could get connected to them. Advertising also leads to a relationship: if you’ve been working with a site for a while you can develop it into a ‘front of mind’ relationship.
There’s also a suggestion that domain-relevance may be build by a brand’s proximity to keywords across the web.
Content should be one-of-a-kind, relevant, helpful, uniquely valuable, likely to spread and have a great UX. The issue we face today is that there’s massive growth in content output, but 95% is still getting no shares or links. The ‘useful’ content is not keeping pace.
Imitate campaigns that worked and get insights from tools like BuzzSumo. Generally text performs worse than visual content, which in turns performs worse than interactive content. Finally, consider paid amplification as it can help any piece of content.
We only have some much time and money, so how do where choose where to invest? Unfortunately, the easier something is to measure the less opportunity there is. Rand explained how ‘serendipity’ and chance is hard to track and attribute but can lead to big results, so the only way to scale this approach is to increase the number of potential opportunities for it. If you’re uniquely good at one thing, over-invest and pump as much as you can into it. A higher barrier to entry mean greater opportunity.
Grammar nazi review: What’s that you say? More inconsistent and every-word capitalisation? You guessed it. I also spied an ampersand out of place mid-sentence. He gets away with it as he’s a fantastic speaker.
