Phew; a packed schedule and a very warm presentation room for SearchLove today. Some great speakers and some surprisingly good food, but 9am till 6pm of presentations does require sustenance and sugar to get you through…
The repeated messages were: delight your users, use data to understand your market, and think mobile first.
As I’m an annoying critic of everything (it’s not big-headed; I know what I do is probably riddled with mistakes) I couldn’t help but grade the presentations for their grammar and overall performance. Feel free to pick holes in my hastily uploaded and not-at-all-spell-checked notes in return – I’d love to hear your comments.
Existential crisis management – @hannah_bo_bannah
SEO is still not dead, apparently, but it’s now involving a lot more brand building. Problem: there’s a lot of brand apathy. “Customers would not care if 92% of brands ceased to exist” – building a brand is tough, but doing it on SEO retainer budgets is even harder!
A company is not a brand. A brand ‘places indelibly’ – it leaves a lasting impression. That’s what you’re looking to do (and it should be favourable!). So SEOs are helping companies to make their brands meaningful.
Three meaningful things brands do:
- Find opportunities to delight customers. That means actually responding to and interacting with your audience, but don’t cross the line of acceptable informality and edginess! Is it higher risk, higher reward though?
- Give people the opportunity to define themselves to others – basically trying to shape another person’s impression of ‘me’. Brands can give opportunities for this, and create things that help people look good by sharing.
- Stand for something beyond their products and services – taking a stand that could even hurt your business can create a lasting impression.
Hannah showed a nice ad from Nike that captured a few of the points. Is it so easy if you don’t have Nike-level budgets? Hannah accepts that it’s more about giving creative teams the freedom to run with concepts and potentially take risks.
Distilled spend at least 60 hours manually outreaching and promoting content. That’s another reminder that it’s never just about the content; it’s got to get out to the right audiences.
Grammar nazi review: Far too many ampersands for my liking, and used ‘inferred’ when she meant ‘implied’. Overall a fair performance.
The Landing Page Manifesto – @OliGardner (Unbounce)
Landing page optimisation is about designed post-click experiences that people want to take part in. If online is a hotdog eating competition, optimisation is making hotdogs easier to eat.
Think about attention ratios: how many things/messages are there on the page that can be done/clicked? Get that attention ratio down to 1:1 – the thing you want them to do.
DO:
- Keep the conversion ‘scent’ on everything. Maintain messages, branding and headline topics – it’s context for the user.
- Embrace ‘friction’ (forces that oppose an action). Design experiences to acquire your ideal customer.
- Use ‘go’ words – they place positive thoughts that aid conversion. Basically say things that people want.
- Be obvious. Less can be more – fewer testimonials, for example, can be more likely to be read; higher believability.
- Be obvious with design. Use directional cues (basically like arrows) – tell people where to go. Test contrast for calls-to-action (CTAs).
- Never start a marketing campaign without a dedicated landing page.
DON’T:
- Design like wikipedia – we’re interested in design for conversions not information! “Stop being useful”.
- Don’t dump your company problems on users – like CAPTCHAs for spam protection.
- Don’t use image sliders.
- Don’t use language that’s confusing/not clear for your audience. Tip: try removing the design of your landing page and read it with only the text content. “What is this page about?” – ask some people on usabilityhub. Then “What would happen when you click on the button” and “Is this credible/trustworthy?”.
- Don’t put bad thoughts in people’s heads – if you mention “We’ll never spam” then people are suddenly thinking about spam! They’re stop words.
- Don’t bullshit – if testimonials are real/true, then don’t use them. Social proof can be bullshit, as it’s subjective to each user.
- Don’t use crappy CTA text – it should be ‘I want to ____’.
Clarity beats cleverness almost all the time.
Grammar nazi review: Very stylistic with few words, so dodged any potential issues.
Becoming a digital superhero – Ade Lewis (Teapot Creative)
Why do some online marketers go home happy every day with happy clients, even if they have less knowledge?
Happy clients mean you go home happy too. The message is get shit done.
Ade says try to understand your clients. Know them, then get them to like and trust you to get more referral recommendations! Helps you avoid taking on ‘problem’ clients (apparently hairy truckers named Dave are to be avoided).
Don’t continue to work for toxic, time-sink clients. Sack them to get back your profitability and job satisfaction.
- Copy stuff: use RSS feeds and tag stuff up some you can find interesting ideas when you need them.
- Learn from others: Don’t get stuck in a way of doing things. Learn from the success of others.
- Do a competitor content audit: Crawl best competitors, run through URLprofiler and collect data about their best content (shares and links). Also run it through Topsy to see who’s influential that’s shared it.
He really pushed that we should start delivering real change for clients by getting shit done; he suggests that’s what makes you a digital superhero. I just end up thinking of a pixelated superman to be honest.
To help with getting stuff done, remove your personal distractions and then learn to focus. Track your time so you can ensure you’re doing your job – not somebody else’s – and the stuff you enjoy.
Ade really pushed some cool tools from linkprofiler. It’s a great tool but it felt quite pushy and almost like a sales pitch for linkprofiler (he admitted that he’s pals with the owners).
Grammar nazi review: Erratic formatting and initial capitals, added to incorrect use of apostrophes, was off-putting. The flow of the talk was a little hard to follow.
Are you Google Analytics reports pretty little liars? – @AnnieCushing (Annielytics)
Be careful of ‘nebulous user data’
An Analytics ‘user’ report is actually on ‘clients’ – it can’t track across platforms or clients used by the same person e.g. at work and at home, or on mobile on the bus.
These metrics are a steaming pile of bull:
- Users (unique visitors)
- New users
- Days since last session
- % new sessions
- Count of sessions
- User type (new vs returning)
If a CEO actually believes those metrics, then you could be held accountable for placing trust in this reporting.
GA userID was released to address this issue and track users across multiple devices, however it’s clear that very few sites will really be able to take advantage of this (because you need users to be logged in, a CRM-attributed ID of some sort, a new Analytics view etc). I don’t think we can be too harsh on Google for this, as in all honesty tracking across devices is extremely hard. If you do implement it, however, you’ll see unique visitors shrink significantly and only see data from logged-in users.
Don’t ignore logged-in users
Use custom variables or custom dimensions to track these VIPs – stalk them!
Some things custom dimensions are good for:
- Add customer id as a dimension to see what they do on-site. Find out the most valuable users then segment them and email market to them specially with promotions etc.
- Track logged in users
- Content success by author
- Content success by page category
- Content success by publication date
- Performance by gender, age etc.
- Membership level
- Number of help articles viewed
- Complaints
Missing metrics that matter
Bring in your own custom metrics and data. People want to see their actual profits from the website. Try importing:
- marketing email opens
- goods costs
- profit
- margin
- population in your target metros/cities
Basically if it’s data that’s out in the world somewhere, you can grab it / scrape it and put it in Analytics.
What does the company care about? Make sure you’re reporting on it!
Heck, an airline could take data directly from their check-in kiosk and put it into Analytics to see how their flyers behave and market to them specifically. Nice eh?
Campaign tagging
Labels matter! Think about your campaign tagging (medium, source, campaign etc). It’s so important to get medium correct; otherwise it will ruin a whole load of reports. Use the Google standards – so if you’re paying for it mark it as ‘cpc’.
Don’t ever put tagging on internal links, obviously. That’s madness.
Third-party services
Yes they add cool functionality, but they will often destroy your data or steal it, never to return. Don’t end up giving all your purchase data to PayPal or some other service.
So, either a) keep people on your site or b) set up decent cross-domain tracking. Hopefully you’re fairly familiar with cross-domain tracking, but making sure it’s set up correctly can sometimes be tricky. Also remember to set up autoLink, as this will help Google Analytics work correctly on its own.
KissMetrics is pretty well set up in Annie’s opinion, as it isn’t dependent on logging in for userIDs.
Grammar nazi review: Mostly charts (well we are talking about data) and no issues, but there were some nice examples of correct hyphenation. Good work. Annie is a very engaging, passionate speaker so top marks (but don’t surprise her – she hates that).
Measurement behind integration marketing strategies – Mackenzie Fogelson
No online marketing works in isolation.
People want a shortcut to building a market-leading brand, but…there isn’t one. Real marketing takes longer, but gets you there with the correct foundation. However, you never get a few years! People want results now, so you’ve got to show value all the way along.
Integrated is building a better, real business, from the inside out, and helping to improve the whole business. It builds real relationships and community, and that then leads to revenue. What’s the core message that applies to the real target audience?
Build a better business > attract a stronger community > grow your audience/brand
Integrated marketing has to be based on your overall organisation goals – create a strategy from those. Then what tactics and deliverables come from this? Then analyse, critique and repeat (Mackenzie recommends 90 day cycles).
You can’t always directly measure everything that you do, and because of this you must always start with the goals. These can be:
- visionary
- business/brand
- campaign
Break the visionary goals down into your campaign, or action, goals.
Measurement is tricky, as everything is working together. There are so many metrics that they can’t map to the tactics being used. You need to communicate the power of integration and how it drives results as a whole. Look at year-on-year or over long-term time periods, and emphasise the progress that the business and brand has made. Choose your KPIs well: how do you tell them how their brand has improved? Social media engagement, returning visitors. Easy relationship building metrics just don’t exist godammit! Focus on your key targets, as that relates directly to your business goals.
Helps avoid “We don’t rank for X” or “We want more social followers”. These comments don’t relate directly to agreed goals, so can be ignored for now and considered as a bi-product of the integrated activity.
Don’t measure (or at least don’t focus on) anything that doesn’t lead to actions or changes. Move issues to a worksheet or workflow (something like Google Docs) and allocate so that nothing is missed in the short term. Does it need to be addressed in a 90-day timeframe? Does it align with their goals? Do these actions justify an increase in budget?
Grammar nazi review: Fonts and styles meant capitalisation was all over the place, but it was stylistic so she easily got away with it.
Marketing in your sleep – Will Reynolds
Will helped us learn from his screwups. Better his than ours eh?
He says don’t be a slave to the content calendar, and regularly posting crap. Content doesn’t have to be flashy and overdesigned – users want “the answer”. They want unique insight. So, ask your customers (or audience) what they want; what they want answered.
To be lazy in your marketing, do one thing really well.
Sometimes doing one thing well can have big benefits far in excess of links, as it leads to ongoing traffic generation from a relatively small initial investment. You could get the same impact as the initial spike over and over again while not doing any work!
Big content is a barrier to entry – @dr_pete. So that’s the way to differentiate and distance you from your competitors; but it hurts when it fails. How do you know if something’s working or not? When do you abandon it if it’s not working?
- Do I have a good data set?
- Is this problem worth solving?
- Can I use it daily?
- Was it over-designed?
Will says give yourself a 20% ‘oops’ budget that covers any ‘oh my lord this didn’t work’ moments.
Don’t think about sales funnels; think about frustrations. Try inspectlet, as it can make analytics more visual, video-based and understandable. You can even see pinch-and-zoom from mobile! There’s a hesitation report for form analytics, so it’ll tell you which boxes people are having trouble with.
A note: don’t leave money on the table because you’re chasing links. Get business! Get revenue! Any links and brand benefits are just a bonus, and give you that nice long-tail ‘doing your job while you’re asleep’ effect.
Expert tip: You can target people with ads in gmail who are with your competitors, if you know your competitor’s email title! You could target them with retargeting and ads after a long delay window when they’re fed up with their current supplier.
Expert tip 2: Try advertising against people searching ‘cancel X service’ queries. Then also serve them display ads on sites and pages that rank for these queries! You could even improve the displayed offer in retargeting the longer they go without signing up.
Grammar nazi report: No notable issues from Will; he’s a smooth performer and very engaging.
Data driven SEO – @dsottimano (Distilled)
Meaningful SEO data is hard to come by and sure as heck Google isn’t going to give it to us. What we consider good is not what Google considers good, and they’re very vague about what they want to see.
The thing is that trust factors vary from user to user, and ‘best practice’ doesn’t really mean much. Every instance needs testing.
Some tools
- SEMrush: get screenshots of historical SERPs, which is pretty useful for researching ‘blips’.
- NerdyData: search by = source code from across the web.
- Peek by LinkRisk: see the risk of domains before you play with them.
- Delimit: does something with CSVs – I missed the point. Sorry!
- Import.io: an great, easy-to-use scraper.
- Scrapy Cloud: an incredibly fast scraper.
Predictive analytics
Dave asked why we, as people, are analysing huge amounts of machine data. Can we get machines to do it for us? Can we predict things like how successful a blog post might be by machine analysing data? Turns out we can’t at the moment; it’s incredibly hard. And complicated.
So he suggests using bigml, which he used to analyse some of his blog posts. He found that low word count meant a crappy post, but none of the other data was amazing or gave direct insight (by his own admission). It did, however, show a lot of potential for understanding what ‘great content’ actually is. Machine learning could be the future.
We should all start by labelling our content, so that at a later date we can properly analyse it. That’s a good tip – start using your meta keywords (that don’t do anything for SEO anyways) to tag your output.
Bin best practice – test it instead. Prove it with data or it didn’t happen.
Grammar nazi report: Pretty simple and minimalist slides from Dave so little opportunity for error, but there were one or two capitalisation and punctuation mishaps. He occasionally used ‘less’ when he should have said ‘fewer’. Overall pretty strong.
Bulletproofing local SEO in 2015 – David Mimh
Google’s vision of mobile and their desired user journey is very different to the standard marketer’s approach. To Google it’s moving towards predictive search (based on location, habits) and then push other mobile experiences such as payment and reviewing businesses.Display of knowledge graph results is also becoming visual to suit people on the move (and wearing Google Glass?).
So how do online marketers report on success when everybody sees different results and lots of people will never even leave Google? What’s your metric when you’ve no visitors to your website? We all need to start thinking about this changing search world.
The US pigeon update was discussed; it’s basically a tweak to the way local is ranked and considered. Most notably it reduced the number of seven-result ‘map packs’ in Google searches, and Google has gotten better at associating queries with relevant local brands. Google has also gotten better at identifying where you’re searching for and results are now more based upon the company’s distance from the searcher rather than the centroid (city/town centre). Directories also (somewhat unjustly) appear to be doing better as the algorithm favours these and bigger brands.
- Click-through rate is becoming a more important factor in local.
- Consider site architecture (traditional directory structure to location pages, potentially through a store locator), fully build out a page for each store (name, address, phone, answer top questions asked by customers at each branch).
- Manage structured data and consistent citations (use the major local data providers – search for ‘local search ecosystem).
- Get reviews! They increase click-through and massively increase trust to note but two reasons. Google+ but also the directories and review sites that rank well in your vertical.
- Ensure you’ve got great photos on your business page (especially if you’re a restaurant).
- Stick to big brands – be a ‘barnacle’ and optimise listings on directory sites that might outrank you. Build links to a Yelp profile for example.
Grammar nazi review: Kept it visual, so bullet dodged! He certainly knows his stuff about local and that shone through.
The threat of mobile – Will Critchlow
Desktop search volumes are falling. Will asks: are we doomed as SEOs?
Well mobile growth is fairly startling; the scale of the growth of Android is vast even compared to something monumental like the iPod. This year there will be more mobile searches than desktop, and it’s starting to cannibalise desktop search. The problem is that conversion rates seem low on mobile, and we’re not very good at measuring mobile traffic. Multiple screens/devices? We can’t really track that. We, as online marketers, need to embrace mobile.
Apart from making your site responsive/mobile-friendly, what else can we do?
Email marketing: email marketing is now largely mobile marketing, as huge numbers of emails are opened via mobile.
Content marketing: it looks like mobile consumption of content could hit 50% in some cases pretty soon, and a lot of that is due to massive mobile growth on Facebook. Facebook use on mobile will soon rise past 50% of sessions. So, content should be made ‘mobile first’ – build content for the mobile audience primarily. “Don’t build differently; build different things”.
Expert tip: There’s a new bit in Chrome ‘inspect element’ – click on the little mobile phone and you can emulate different devices and networks. That’s actually incredibly useful to quickly see how content appears on mobile.
Ignoring mobile is so dangerous for content marketing and virality. If you’re eliminating half of your audience the chances of achieving a ‘snowball effect’ is massively reduced. Don’t shoot yourself in the foot: build different things that think about mobile first. Google has even revealed that it develops mobile first and rolls out to desktop subsequently.
A worry is that app search is decades behind – there’s no long-tail and it’s mostly ‘branded’ searches. Instead it’s seemingly becoming integrated into regular Google search in various guises. Want to rank an app? Don’t forget the following:
- Get basic app indexing (enable deep links – so a search result can open directly into an app)
- Map those deep links in app pages to equivalent normal site pages (using rel=alternate, schema.org, xml sitemap) – remember the first click must be free i.e. a user must be able to see the page without being logged in.
- Use Google’s tool to test ‘how would this open using an app?’
It seems like the app indexing API is using user activity to rank; that’s interesting and one to keep an eye on, as we’re close to Android ranking apps you have installed much higher.
We’re moving past mobile however; soon everything will have a URL. Everything in life (the internet of things) can be a Google card; an entity on the internet. How do we deal with this? A question for another day it seems…
Grammar nazi review: I suppose using all-CAPS reduces that chance for errors, but there was still the standard inconsistent punctuation typical of presentations. Nothing offensive, and his overall performance was very good as always.
