In 2006, before iPhones and Androids, a Google document mentioned the search engine uses “over 200 signals” for ranking. There are probably thousands of ranking signals now and we can’t pretend to know what they are. What we do know is how much of an impact optimising for the most significant ranking factors has – and how easy it was to get that done in the real world.
We work with almost 100 businesses internationally, tracking on average 15 competitors and 500 keywords per website. We took the 2015 Moz Ranking Factors study as a starting point (since we contributed to that anyway) and assessed what has become more or less important; what’s more important for the enterprise brands we work with; and what we can actually get done. If a change is easier for you to make – congratulations! You’ve got a competitive advantage.
View the study in full and rank factors by their ease and importance using this interactive asset below:
Top 10 ranking factors you should be optimising for in 2018
- Page-level keyword and content based metrics – If your page isn’t relevant for the query – but more importantly doesn’t answer the user’s intent – you don’t rank. Most businesses from SME-size upwards use a content management system which should make this easier. It’s not all about publishing more – the biggest, quickest wins come from optimising what you’ve got.
- Dwell time or long click metrics – When you put a page live, the question you should ask yourself is: “how can we absolutely guarantee that the searcher won’t click the back button?” You know more about your business and products than anyone and it’s time to demonstrate.
- Existence/quality of verified real-world business info – The combined traffic driving power of voice search, the local pack, Google Maps and the Knowledge Graph is immense, and it relies on the accuracy of the information its being fed. You should absolutely clean and own your data across the web. Optimisation of WikiData and location data are not only hugely important, but don’t often require development resource or approval from compliance.
- Use of responsive design and/or mobile optimised – This is not only a prerequisite of ranking on a mobile device, it’s also getting easier and easier to get buy-in. Brands without mobile websites in 2018 are at the tail-end of the “laggards” phase and it’s easy to quantify how much money you’re losing out on without a mobile site.
- Uniqueness of content across the whole site – De-duplication of content across the website speeds everything up, prevents cannibalisation and makes better use of crawl budget. Most modern CMS systems feature some way to deal with this issue. You can even take some shortcuts in Search Console.
- Page is mobile friendly – Mobile-friendly is a rare example of Google telling businesses, in a developer-queue friendly format, what they need to do to pass – and giving the seal of approval when the standards have been met.
- Uniqueness of content on the page – Have the best possible answer on the web, whatever the question being asked. Research every possible intent and constantly improve your best performing pages.
- Page load speed – The fastest page usually wins. Again, Google gives us a to-do list – and compares us against the rest of our industries, which has made buy-in a little easier to get across our clients, but advancements in technical SEO are accelerating. Content delivery networks (CDNs) are the norm and site speed should be at the top of everyone’s technical agenda.
- Quantity of searches for this keyword + specific brand name, URL or domain – We’ve proved this is a thing, but this one is a little bittersweet. Searchers are doing this partially because it’s easier to search Google for pages on your website than use your navigation or your own internal search. So we’d much rather you optimise those things – and so would your users.
- Relative CTR from Google SERPs to the page for the keyword – Optimising for click through rate (CTR) is so easy. Schema mark-up is relatively straightforward to implement, giving us better looking listings. We constantly test title tags and meta descriptions too, which have often been branded as “not a ranking factor” and have therefore been ignored (budget constraints usually). If we can get a better CTR in second than the result in first, then rankings don’t even matter – but the results will usually swap over time anyway.
The elephant in the room
- Links – We are absolutely confident that links are the single biggest ranking factor. But link building is hard. It’s hard to get people to link to you. It’s hard to pinpoint the exact value of an individual link. All our metrics are bullsh*t. Most websites benefit more from links than from any other single ranking factor, but it will suck up a disproportionate amount of time and resource if you don’t manage it carefully.
Methodology
The Moz Ranking Factors 2015 study was our starting point (it’s the most recent version at the time of writing). 150 very clever people (including our own Tim Grice) answered Moz’ survey to put it together and who are we to argue?
…but the study is two years old, and although we don’t personally know every single contributor, we do know that we work with very different businesses in many cases, in some of the most competitive industries worldwide.
So, using the aggregated scores from the Moz panel as a starting point, we asked each member of our SEO team to score the same ranking factor out of 10 for importance. The 150-strong Moz panel is one vote – our SEO Strategists get one vote each, in isolation. Collectively, we’ve worked with hundreds of large websites – at Branded3 and at previous agencies and in-house roles for some of our recent hires – each with hundreds of competitors we’re keeping a close eye on.
Then we asked our SEO team: in your experience, how difficult is it to optimise for these ranking factors?
We’re taking into consideration:
- How our clients’ SEO teams are working with other internal teams such as development, PR and compliance
- How our clients’ teams work with their other agencies
- How far our clients’ budgets typically stretch
- How far down our clients’ lists of priorities SEO activity tends to be.
How we use this study
If you disagree we’d love you to tell us on Twitter or email me directly…
…but really, if your experience says that optimising for a ranking factor is easier than we do, congratulations! You have a competitive advantage. If we all agree that something is a significant SEO ranking factor but you think it’s easy to tick off your list – and we’re telling you everyone else struggles to get it done – it should absolutely be a priority when you’re putting together your SEO strategy.
…especially if you’re currently focused on something that might not be so important.
Tactics that are extremely important but also extremely difficult to optimise for tend to be our long-term strategies. If we’ve previously been looking at things that are difficult – but not that important – we’ll try to move the conversation onto something we think will have more of an impact.
So without further ado…
Our full 2018 Google ranking factors matrix
Top 10 most-important Google ranking factors
Top 10 easiest things to implement
10 things we’re confident aren’t important for rankings
Variance
The 10 things that are more important in our experience:
- HTTPS/SSL – The difference here is definitely 2018 (us) vs. 2015 (Moz). Though Google had already claimed that HTTPS was a ranking factor before Moz published its survey results, it was largely untested, with few big sites making the switch. Google has since made several changes increasing the overall importance of HTTPS.
- Dwell time and long click metrics – We’re collectively 100% convinced time to long click is a huge ranking signal after numerous tests circa 2013-14 – nowadays we try to optimise our clients’ sites to better retain visitors as part of every strategy.
- Keywords associated with domains through entity association – It’s why AutoTrader will always rank for “used cars”, for example. The brand is synonymous with the keyword.
- Keyword is present in the root domain name – Exact match domains work, even though they shouldn’t. Some of the businesses we work with have a couple of core keywords where EMDs are extremely difficult to dislodge, even though they look like spam sites.
- Domain-level, keyword agnostic features – Technical SEO has accelerated since 2015 and the industry has collectively bought into tactics like site speed optimisation (thankfully). Links are harder to come by, which may be part of the story, but actually new technologies like AMP are providing us with a new competitive advantage every month, if we can get them implemented.
- Quantity of searches for this keyword + specific brand name, URL or domain name – See above.
- Responsive design and/or mobile optimised – Again this is probably 2015 vs. 2018. Mobile optimisation is becoming more important as Google rolls towards its mobile-first index, so this should be high on your agenda.
- Page load speed – The faster page usually wins. ‘nuff said.
The 10 things that are less important in our experience:
- Bounce rate – Google doesn’t use bounce rate, although Bounce Rate as a metric in Google Analytics is often a good indicator of more significant metrics like dwell time (long clicks).
- Link velocity of the page and page-level link metrics – Internal pages don’t naturally acquire links very often and it’s unreasonable to reward pages that do because there’s almost certainly something going on to make it happen.
- Topical relevance of linking domains – Again, we’ve written about niche sites Our link acquisition strategies tend to revolve around national press and high authority sites who tend to write about everything. Niche blogs can be good to drive traffic but aren’t usually as good for SEO.
- Domain is associated with high-authority authors – Author Rank used to be (rumoured to be) a thing and now it isn’t. We’ve moved on in the last two years.
- Anchor text – Causes penalties. It’s also nearly impossible to get good links with anchor text, so unless you want to get branded anchor text you’re going to be disappointed.
- Position/context of inbound link – Links in content do fare a little better than sidebars, footers etc. – but does anyone do that anymore?
- Sentiment – We get great results from regular links on the Daily Mail, where the sentiment is always negative, regardless of what we’ve done.
- TrustRank – Is a bit more complicated than metrics like MozTrust or Trust Flow would suggest, and we don’t believe it’s based on distance from a trusted set of pages/sites – rather that sites in X industry are expected to have links from a set of specific sites to prove they’re legitimate.
What should you do about this?
We don’t think Google uses totally different ranking factors for big brands and competitive queries so what you need to optimise for depends entirely on what you’re able to do. It’s not helpful to say that Google ranks a website based on a factor that we can do very little about (e.g. Rankbrain) – your starting point should be to decide what’s disproportionately easy for you to do. Do you have less red tape than brands in your sector? Do you have a bigger development team? It’s only a factor in whether you rank if you’ve managed to do something about it.
