They say the first step is admitting you have a problem, so here goes: analytics tools have a real problem – they kinda suck.
Let me explain.
In the late 90’s and early… noughties (ugh, worst decade nickname ever), we were all sold on the glittering potential of the near-infinite measurability of the internet – through the magic of the browser, we were told, we could track and know everything about our customers. Surely with this advanced new medium, we’d be able to get far better insights into our customers’ behavior, beyond anything that simple Nielsen ratings and print circulation numbers had given us before, right?
Well, one part of this utopian dream came true: yeah, we’re tracking everything, but now we’re drowning in meaningless data. If you were to ask your team “How’s the website doing today?”, you’d better get cozy for the next few hours, while they tell you 10 conflicting stories of how today is both the best performance day ever, yet the sky is also simultaneously falling.
Seriously, I want you to physically nod your head if the following sounds familiar:
“Traffic is way up from Twitter (but we don’t know who tweeted us), and our avg. time on site is up about 11%, because we just published an image gallery post – but we’re not sure… The bounce rate is low, except on our pricing page, where it’s 24% higher, but that’s probably just because it’s Labor Day weekend in the US, and Apple just launched a new iPhone… ¯\_(ツ)_/¯” ~ Your marketing team
How is it in any way acceptable that we have all this data at our fingertips, yet often the best conclusions only come after several hours spent hopping between 3 different tools and spreadsheets? Are analytics tools truly doing their job, if users have to ask “Well, what does that actually mean?”, after using them? No! Their job is to analyze that data – it’s in their damn name.
“There are too many spreadsheets and Trello boards still involved in a growth marketer’s workflow.” ~ Ty Magnin, Growth Marketer at Appcues
In this post, I’m not just going to gripe about analytics (as fun as that would be) – I’m also going to talk about the insane potential of better web analytics tools to improve the entire internet experience for everyone, and the challenges we all face in realizing that potential.
So pick a section & dive in, why don’tcha?
Why would better analytics be so beneficial?
Why the current analytics reality sucks
What does better analytics look like?
Who’s thinking about analytics the right way?
Why would better analytics be so beneficial?
Ok, so I’ve already described what using most analytics tools is currently like today, and you’re probably thinking “well, what’s the big deal”? As Simon Sinek says, “Always start with why” – so imagine for a minute what interacting with your analytics suite could be like.
What if…
…you were proactively alerted whenever something significant was happening on your site?
…reviewing your website’s performance took just a couple of minutes, and resulted in clear decisions, and defined next steps for you and your team?
…you always knew whether you’re on-track with your goals, and when you need to step things up?
…you were told in plain english exactly where the best opportunities are to improve your key metrics?
And what if this could all happen in real time?
Life could be a dream…
Now consider how much daily life would be improved for the average marketer or product manager, if even some of the following things were true:
- No more weekly performance meetings. Ever. You’d never again need to go to that goddamn weekly meeting – everyone would just know what’s up.
- You’d focus more on impactful work – Instead of having to regurgitate stats in an interpretable format for whomever signs your checks, you’d be able to just get on with the meaningful work of making those numbers better.
- Content quality would go up everywhere – If we truly knew what content was performing the best, we’d all be making tons more awesome stuff, period.
- No alarms, and no surprises – Rather than feel that pit of fear in your stomach when you go to log into your analytics dashboard, you’d know exactly where you stand vis-a-vis your monthly goals. Compare that to the status quo, where your traffic could triple overnight, thanks to a tweet from some influencer, and you wouldn’t hear about it until 2 weeks later, when you stumble upon the spike in your traffic reports.
Whoa, I want to go to there… So Why do today’s analytics tools kinda suck?
While it’s easy to bash the state of analytics today, it’s somewhat harder for most people to empathize with the humongous obstacles faced by startups in the analytics business today. And while I’ve done a bunch of griping so far, I’m also here to attest to the fact that the challenges are real, and they are steep.
There’s a high technical barrier to entry
Think about how many individual bits of data are currently languishing in your Google Analytics account. Think about how quickly that body of data grows every day, and imagine trying to record, check, crunch and display it all for thousands of websites simultaneously. Yeah, it requires a significant backend architecture to build and scale an analytics product.
We fundamentally misunderstand the meaning of the word “Insight”
“Your traffic is down 17% today” — I can’t do anything with this info, except ask for more info. This is what most analytics tools will tell you is an insight, but this is not an insight.
“Your traffic is down 17% today, because 3 of your top referrers removed links to your pricing page yesterday.” — Now this is a frickin’ insight! I can actually do something useful with this info. Sadly, I see very few analytics products approaching this depth of analysis.
Analytics tools need to stop reporting raw data by default, or showing simple higher/lower comparisons & calling it an “insight”. Tell me about significant movements up and down in my key metrics, give me my bests, averages and worsts in the data sets you’re analyzing. I need a barometer and context in order to know what’s going on – simply knowing the current temperature is 112 degrees, for instance, does little to spur me into action, unless I already know that the average high is 77 degrees, (and that above 115, my body will shut down).
The myth of the Google Analytics standard
Google Analytics is the perceived gold standard of web analytics, but we’re all being held hostage by the false belief that Google Analytics’ numbers are sacrosanct. Most people actually have less trust in a smaller paid analytics product, than a free one from a company that earns money from selling our eyeballs to advertisers. #definitionofirony
Big-data analysts probably sleep on big piles of money every night
Sifting through all this data for useful patterns and insights to share with your users is a tricky job that demands a very particular set of skills – and with those skills comes a commensurate price tag. Speaking from my own experience building Filament.io, those skills are critical for finding the types of highly-actionable insights that we’re all crying out for, but they don’t come cheap.
People don’t ask for things they don’t know are possible
Data visualization techniques and the front-end capabilities of the internet have progressed by inhumanly huge bounds in the last few years – but look at the design of most analytics tools, and you’d be forgiven for thinking the pie chart is the single greatest innovation in the industry since 1998.
What does a forward-thinking analytics platform look like?
AKA, the part of the rant where I lay out my wishlist for an absolutely ideal (I think) web analytics product – one that I’d want to work with. Most of these features aren’t new, there probably are tools that offer some of them – but very few that unify these things together in just the right package.
- Voice activated – don’t laugh. If I can tell my phone to automatically handle my pizza delivery, I should be able to ask it how my website is doing, and receive a semi-intelligent answer.
- Natural language data interpretation – tell me what’s working, and why. Then tell me what’s failing and why. And do it all in natural language responses.
- Automated pattern/correlation recognition – I want a system that’s always watching for emergent behavior patterns that lead to good things, like purchases or subscriptions, and then tells me about them.
- Realtime push notifications of significant activity changes, like good new referrers, spikes in purchase activity & traffic, etc. – really, I should barely ever have to log in.
- Intuitive goal-setting & projections – when am I on-track, and when am I lagging?
- A user-friendly mobile app interface – I want to stay connected with my site at all times, in all places. I don’t want to be staring at charts on my smartphone.
- Integrated social stats – tell me which influencers are doing me a solid and sending me good traffic, and show me where the conversation about me is happening.
Who’s thinking about Analytics the right way?
While I might be the only one throwing a textual temper tantrum about this, there are actually a bunch of super-smart folks already poking around in this analytics goldmine and trying to realize its true potential:
- Michael Ruminatsau and Mika Melchenka, Narrative BI
- Jeff Weisbein,
Know Your Analytics (KYA)(RIP) 🙁 - Christopher Penn – The Next Evolution of Analytics
- Shout-out to the new
Filament.ioteam (RIP) 🙁 - Ash Read, Postreach
- Chuck Reynolds,
Levers(RIP) 🙁
The fact that half of the analytics tools created by this distinguished group of leading thinkers have already disappeared should tell you something about the difficulty of this problem.
What gives you the right to be so hoity-toity about this?
Believe it or not, even though I nearly failed high school math, I do actually love web analytics – or rather, I love the things they can tell us. But I’ll tell you what: we’re all depriving ourselves of massive benefits by not demanding more from our analytics tools – we’re missing out on potential progress on multiple fronts, from better quality content across the board, to more accurately targeted and relevant marketing that wastes less of everyone’s time.
And that makes me a little grumpy, because no-one wants to create a sucky website – we just don’t know how to reliably make a good one, because the tools we use to measure its performance place too much burden on us – the squishy, forgetful, inaccurate humans – to derive high-level meaning from the data. Building better analytics tools would go a long way towards driving an immense improvement in the overall internet experience.
PS It’s my secret hope is that this rant can annoy someone out there enough to decide to shut me up by delivering a superlative analytics platform, and we can all live happily ever after, (praise Jeebus).
Stuck in Analytics Hell? Let me help
Digging into your SaaS metrics for useful insights is all well & good, but sometimes you just need it done for you, while you’re working on other stuff.
GrowthLook is a quick-turnaround, flat-fee SaaS metrics analysis service I offer, that gives you solid, actionable revenue growth recommendations tailored specifically for your business, in 7 days or less. Check it out!
3 comments
Comments are closed.