BrightInfo Review: Grab anonymous visitors before they bugger off

You know those pop-ups that, well, pop up and make you an offer when you’re about to leave a website?

Well they always bothered me. They feel kind of desperate or crassly self-serving, like the McDonald’s order-takers who ask if you want fries with that.

But two things kept me from railing against them:

1) I do sometimes ‘engage’ with them.
Sometimes, the offer they make is appealing enough that I click. So the marketer in me (also known as ‘me’) files that away under ‘stuff that just might work for our clients’.

2) A HUGE proportion of web visitors come to us once, don’t leave a name, and never come back.
And the data I’ve seen suggests this is typical. To which I can only say, “OUCH! Ouch-ouch-ouch-ouch-ouch-OUCH.”.  All that time, money and effort to get people to your site and they wander off FOREVER? Without so much as an email address or a box of chocolates? The marketer in me (me again) thinks about this and cries bitter, blood-tinged tears.

So we figured we’d give this pop-up or ‘exit-triggered promo’ thing a try – just to see how well it works.

We (actually Daniel Smulevich, one of our clever analysts who has since left Velocity for a better-paying but ultimately soul-destroying job, crack addiction, homelessness and multiple cat ownership) chose to work with BrightInfo for our experiment.  (If you want to see how these things work, go to their site and try to leave).

This is what they say of themselves:

“BrightInfo discovers anonymous visitor’s intent in real-time and dynamically offers your most relevant content – from white-papers to videos to case studies.”

And that’s why we chose them for our experiment. Because if we were going to risk annoying our visitors, we wanted to do it intelligently (which, we hope, means annoying fewer people and annoying them less).

It’s like ad remarketing (which we also hated but tried and now hate significantly less): if it’s done intelligently, it’s less intrusive, more helpful and lot less annoying.

In short, we didn’t want to do stupid things in public: like using static content promos that push the wrong content at the wrong people or shove a piece at someone who’s already read it.

So could we make pop-up “Don’t Go Yet!” appeals less annoying?

And if so, would they actually work?

That’s what we set out to discover.

How it works.

The BrightInfo processor crawled our website, using a mix of methods to identify Key Terms for each page (mostly phrases rather than single words). Geeks might like to know that they use a combination of Natural Language Processing (or ‘magic’), complex regular expressions (say what?) and HTML analysis (sounds clever). Then, to validate the key terms, BrightInfo applies statistical analysis and identifies synonym relations between terms.

After the analysis, BrightInfo automatically matches the promoted content to each web page based on their scoring mechanism.

Then, in real-time, the tool decides what to show to the visitor based on the information extracted by the site processor, the visitor profile and behaviour (entry page, navigation on site, history, time on each page, etc) and statistical/behavioural analysis.

Crucially, the pop-up can be suppressed for low-engaging visitors, so you’re only hitting people who have shown some level of involvement (three blog posts read for instance).

The idea is simple: if you’re going to ask, “Do you want fries with that?” to a prospect on the way out the door, make sure they like fries, haven’t already ordered them and aren’t potato-phobic.

Here’s what it looked like on our site.

When we did the experiment, BrightInfo didn’t let you really design the lightboxes and panels – just specify colour and choose from a limited font menu. They’ve changed that (I think our constant whining may have played a part) and now you can pretty them up a lot more and make them fit your website.

There are two kinds of panel: a sidebar that’s always there on every page you don’t exclude; and an ‘exit widget’ that can be configured to pop up in the centre of any page you choose when the visitor shows signs that they’re about to leave (a quick mouse move to the top of the page is a biggie). According to BrightInfo, it’s not uncommon for people to click five times more on the Exit Widget than the sidebar — but the overall performance is 30-40% of the sidebar because the latter is displayed far more often .

This is a typical sidebar:

It’s tucked in some open space, outside the frame of our page, so it doesn’t overlay any of our content. As you can see, it competes with our own content suggestions – the yellow bar on the right – and tag-driven ‘Related Posts’ beneath every post (out of view in the screen grab above). But those are static — no targeting or intelligence.

Note: this is NOT a pop-up, it’s an always-on panel that, ideally, feels like a natural part of your site. So even if the pop-op makes you nervous, you can go with just the sidebar.

This is the Exit Widget offering one piece of content:

Just before you exit, the whole page goes dark and a single piece of (usually high-value) content is suggested. The one shown is not pretty but BrightInfo has changed that so you can brand it better.

You can apply this pop-up to a single page of your site or every page — or just turn it off. So if the idea of interruption bothers you, just don’t use it.

This is the multi-content Exit Widget:

Offering three pieces with roll-over extracts. The Software will rotate pieces through the lightbox and change lightbox styles to find out what works for each page.

What happened when we gave it a go

The tool was active on our website from 10^th April to 21^st May, 2014 but the data shown below are for the period 14/4 – 18/5 inclusive, to show five full weeks.

The Topline:

Compared to the previous five weeks, Google Analytics showed that our bounce rate decreased by 7.27% and exit rate by 4.44% . For our site, this isn’t life-changing but for a high-volume site it is more than significant.

Some drill-down:

The above chart shows the uplift that the BrightInfo sidebar gave us for the top five content pieces promoted.  The “number of native visits” column counts only the visits to this content from other pages within our site, excluding visits direct from search engines, so we’re comparing how our site performs with and without the widget.

As you can see, BrightInfo did get people to view content they wouldn’t have viewed otherwise. And that’s the whole point.
The percentage increase is a bit misleading as the numbers are small but just on raw views, this looks worth doing — especially considering that these people were on their way out the door.

Upside

• The experiment delivered positive results and let us tailor content recommendations to users (rather than simply to an article or page).

• The longer you leave it on, the better it works – making better recos based on performance (without having to touch it).

• There’s potential to use ‘virtual gating’, (which we didn’t try), so you can capture an email even for content that’s ungated on your site.

• The analytics are good – and getting better.

• You can assign relative importance to certain pages and sets of pages – and exclude pages you don’t want to show the widgets on.

• The Brightinfo team were super-responsive and helpful. Gotta be said.

Downside

• Integration with Google Analytics is done through utm parameters: We disabled this to prevent referral cannibalization (hate that). Brightinfo promptly disabled that feature upon request. This means we had to use Brightinfo’s analytics only. We’d recommend the use of Google Analytics events.

• Some site integration glitches – Because of our custom theme, the widget didn’t automatically show up on the website’s homepage and the blog’s “homepage” – in the first case, that is actually not bad but it would be good to have the option.

• Weird content choices – Sometimes we felt the tool was inexplicably skewed towards certain blog posts.

• Inability to specify content recos on a per-page basis – Apparently now being rolled out, along with other over-ride features.

• You never know how many people you pissed off for every one you converted – this is the unmeasurable downside.

Conclusion

I went in as a skeptic. I never liked these things. But my lurking reservations about that antipathy proved true: they work. And if they’re done intelligently, as I believe the BrightInfo guys are doing, you can push the upside and mitigate the annoyance factor.

In fact, the sidebar deployment isn’t annoying or intrusive at all (though it may compete with other feature boxes on your site).

Would I recommend it to clients? In many cases, yes, I would. Letting anonymous visitors snivel away without at least pitching something relevant to them seems too big a waste.

I might not recommend some of the other pop-up tools I’ve seen — the static versions with no personalisation, or even the rule-based ones (rules get really unwieldy really fast with something like this). But I do like what BrightInfo is doing and I think it’s worth serious consideration.

So why isn’t Velocity using it any more?

We thought you might mention the elephant in the room.

We’re not using it because the upside just isn’t big enough at our traffic levels to justify the gut-feel downside of annoying the proto-friends who come to our precious website. We also have an ethos about not trying too hard to ‘convert’ our visitors, preferring to engage on their terms, but this reflects our specific situation and is maybe the subject of a later blog post.

Yes, it’s a great way to keep long-tail content exposed. Yes it works. And if you’re worried about pop-ups becoming annoying, you can dial them back or turn them off (without losing the sidebar benefits).

If you’re as appalled as I am about the number of people who drift in and out of your website without ever saying hello, you may want to test drive BrightInfo. It’s ridiculously easy to deploy and is the most intelligent way to approach this tactic.