A few years ago, I wrote a monster post. It became so popular that it dwarfed all other pages on my website (even the homepage) – even though I never expected or intended for it to do so. In fact, as I write this, in the past month it received 73% of all hits to my website – so it’s more popular than every other page or post on the blog combined.
As time went on, I had a bit of a dilemma: the content slowly became more and more obsolete. It’s still partly true/helpful even today, but more and more people had been commenting on the post saying that my method no longer worked for them. Obviously I wanted to leave the post up, for both selfless reasons (e.g. as it’s still useful for some people, albeit fewer than before) and selfish reasons (e.g. TRAFFIC), but I felt that the current version of the post was lying to them, so I reluctantly decided to change the title, from an absolute, this-is-how-it-is-now-and-forever style of ‘how to’ headline:
How to Remove Slanderous Google Reviews – A Case Study
…to something suggesting that it was perhaps more time-sensitive and therefore not necessarily helpful any more:
How I Removed A Slanderous Google Review – A Case Study (2013)
It’s very subtle – I basically made it past tense and reiterated the year at the end. However I was genuinely concerned that it would be a massive turn-off to people browsing Google for help on the subject. The vast majority of the traffic that the post gets is via organic search, and I was well aware that the title change would update accordingly in the SERPs… What if someone thought “hey, this is old – I’m not clicking on that…”
In addition to the title change, I added a sort of intro disclaimer section:
I made the change on 3rd November 2015 – nearly 10 months ago. I thought it’d be interesting to delve into the Google Analytics data and to see if the change had had any majorly positive or negative effect on my blog’s traffic levels.
The Google Analytics data: a snapshot
I jumped into Google Analytics’ Behaviour > Site Content > All Pages section,* set the traffic to organic search only, and clicked on the post in question. Then I tinkered the date range from 3rd November 2015 (the date of the change) and the last full day of data (which was 29th August 2016 as I type this), and compared it to the previous period (6th January 2015 to 2nd November 2015). Here’s what I saw:
* I tried Behaviour > Site Content > Landing Pages at first, but ended up some seeing weird data (e.g. ‘average session duration’ times less than a minute, despite much longer ‘average time on page’ stats – which doesn’t make sense), so I’ve gone with All Pages. (If anyone knows why this might be then please do let me know – drop a comment below or tweet me please.)
However I consider this the most throwaway stat of this analysis: correlation does not mean causation, and it could’ve simply been the case that a) more people have searched for the subject in the latter time period, and/or b) the post’s rankings were better during the latter time period. It’s always ranked quite well in Google – for keywords such as "how to remove a google review", "how to delete google reviews" and "how to remove bad reviews from google" – but admittedly I haven’t kept a close enough eye on it to say that it’s ranking better now vs. back then. At least a buttload of people have still been visiting the post and not running away screaming.
This went up (and up = worse) – but only ever so slightly. From 88.23% to 89.16% (an increase of 1.05%). ⬇️
This is where I was expecting things to be a lot worse, with people bouncing off as soon as they realised that the advice may not actually be able to help them, as it’s old and potentially outdated. But to my surprise, the bounce rate remained (mostly) consistent.
It’s a very high bounce rate regardless, but I think that’s often the nature of the beast when it comes to blog content – after all, if someone’s surfing for a solution for a very particular problem, why stick around on the site afterwards? Most people will click back or close their tab/window. A searcher in need of help on removing slanderous Google reviews is unlikely to browse the rest of my site’s offerings, which includes Matt Cutts caricatures and an example of what Watchmen‘s Rorschach might write in his journal if he were an SEO… But you know what? That’s their loss.
Average time on page
This also went the bad direction (down) – but again, not by much. From 15m05s to 14m15s (a decrease of 5.59% or 50 seconds). ⬇️
50 seconds is… not a lot. At least not comparative to 15-ish minutes. It still suggests that the vast majority of searchers who’d clicked-through were interested in reading a good chunk of the post, despite the insistence in my new version of the title/intro that while it helped me way back then, it may not necessarily help them now.
Another observation: number of blog comments
There was also another (non-GA) stat that I considered looking into while researching this: had there been more or fewer blog comments against the post since the title change?
The post is a beast, having amassed a total of 103 comments and pingbacks to date (as I type this) – although admittedly a good number of them are probably from me replying to each of the commenters. But despite the increase in pageviews (see above), was there a difference in blog comment numbers?
From 6th January 2015 to 2nd November 2015, there were 19 blog comments (not including those authored by me). From 3rd November 2015 to 29th August 2016, there were 10. ⬇️
So there were half as many blog comments left by people in the latter period – despite more visitors to the post overall. Admittedly a blog title that suggests that something happened in the past might mean that people are less inclined to comment. For one, they might think that no one will read it – whether the author (me) or other visitors. However I do think that the old correlation/causation argument may also come into play here: I rarely ever see old blog posts or forum threads with either a steady spread of comments over the years or a sudden resurgence later on. As time goes on, old posts generally garner fewer comments. Therefore it might not be the case that the blog post’s title change is to blame here – it could simply be the case that fewer people have wanted to comment on it regardless.
The final verdict…
I was really hoping that there’d be more to report here – something sexy or exciting at least. Like “OMG, my time on page stat plummeted!” But nope, sadly not – the data remained fairly consistent.
But there’s a lesson here. I was really anxious about changing the post’s title. After all, it was performing so well – and had done for such a long time – that I was worried that any minor change could’ve really hurt it. But it didn’t. And you know what? Now I don’t feel like I’m lying to my visitors (unlike before), so I can sleep better at night while knowing that I’m not shooting myself in the foot while doing so.
Of course, that doesn’t mean that you will see the same if you go around changing your posts’/pages’ titles – your mileage may vary, as they say. But at least it helps us to get out of the stigma some people find themselves in (especially when it comes to SEO) that our websites might explode if we change a couple of words.
A sequel (and another test)
I actually have plans for a sequel of the original slanderous Google review post (as teased on Twitter), talking about an updated process which works. Rather than to take down the old post and redirect it to the new one, I actually plan to keep it live but change its rel="canonical" tag to the new post (if you’re new to canonicalisation then read this). So expect a future post detailing the finding of such an action… 😉
The life of an SEO eh? It’s hella exciting.