Back in 2007 (an eternity ago, in both dog- and internet-years) I was running email marketing for a midsize dotcom. Fast Company had recently pronounced the death of email – of course, we all saw how that turned out. But at the time, as today, we were thinking email would be around for a while, and were all looking for ways to make our emails more effective.
Personalized Emails – The Wave of the Future
The biggest thing we had going for us? Personalization. Everyone was talking about it. Apparently, just by adding the recipient’s name to the subject line of an email, you could increase clickthrough rates by 100%!
It seemed crazy…but it worked. Apparently, customers were so over-saturated with generic email newsletters and transactional emails that anything that called their name got their attention. Those of us who tried the technique thought it was amazing, and started blogging about it and generally recommending it to all our email marketing friends.
…but then everyone started doing it. And we marketers started to see the gains taper off. Our previously-boosted metrics began to fall back to their normal levels. Why? Because people got tired of the gimmick.
And that was the real problem: personalizing emails was a gimmick.
Next came dynamic landing pages. Anyone remember them?
It started with Dynamic Keyword Insertion (DKI). For a long time now, PPC marketers have used DKI to make their pay-per-click ads more relevant to Google users searching for a product. (For those of you who don’t know, you accomplish this by giving Google or Yahoo/Microsoft a list of keywords, and if someone searches for any keyword on your keyword list, the PPC engine will dynamically replace your default ad text with the appropriate keyword from your list and serve it up. To the searcher, it makes it seem like the ad is targeted to the exact keywords they searched for.)
And generally, we users – even seasoned marketers – fall for it. If we see a link with the exact words we just searched for, we assume it’s what we’re looking for and click through.
The lesson to be learned? When marketing online, relevance is key.
With the release of Google Website Optimizer, it became even easier to game the relevance system. You could create hundreds (or even thousands) of dynamic landing pages targeted to the keywords that brought users to your site. When someone arrived at your site after searching for a ridiculously long keyword like cloud address book that syncs Google Contacts, every other word and heading on your landing page could read “cloud address book that syncs Google Contacts,” ostensibly convincing the visitor that this was exactly what they were looking for.
Again, the users fell for it – here they were, searching for something, and found a page that was built to address exactly what they were searching for. Down to the letter! It was almost magic.
This technique was a great way to increase conversion rates, and it was also a good way for smaller marketers to build out a variety of landing pages with less effort.
But there were some problems:
First off, when you use a machine to do a human’s work, things get wonky a lot of the time.
And in the case of dynamic landing pages, we marketers saw a lot of that wonkiness. Misspelled words could show up on pages when someone accidentally mis-typed. Content creators would overstuff keywords, which looked unnatural for some searches.
It was a pretty big mess.
Additionally, people got smart. They started to recognize the process. Today, your average internet user can now spot a page full of fake details when they see it. And when they know something is auto-generated, they tend to disengage.
To top it off, Website Optimizer closed down in 2012, which left marketers without a Google-sanctioned tool to keep building out dynamic landing pages by the thousands.
Regardless, both of the above techniques were effective. But both were missing something.
Relevance, Not Gimmicks
The problem with the techniques above was that they didn’t consider the user. I mean, yes, they considered what actions a user might take…but they did it in a way more similar to observing microbiology in a lab than looking at what a real, live, breathing person might do.
Think about it: when we’re selling a product or service, what are we trying to do? We’re trying to meet a need. In order to close a deal, we want our messaging and all interactions with a potential customer to be as relevant to their needs as possible. Which means learning a little more about them – rather than just personalizing an email and generating landing pages based on search keywords.
So how can you do that more effectively? By listening to (and integrating) all the data points you have about a prospect. This can range from choosing the right tools to developing the right workflows.
Here are a few ways:
1) Link feedback to social context.
It goes without saying that you should regularly ask your customers to tell you how they feel about your product. When you solicit that feedback, use the Klout API or FullContact’s digital footprint data to identify what topics they deal in regularly. If you spot a trend that small businesspeople aren’t finding your product to be useful, you know to focus your efforts on that segment for a while.
2) Automate your onboarding workflow with context.
Using social media or other data, you can flag certain people who sign up for your service based on location, job title, or influencer status. Our favorite example of this is Maptia, who uses social media data to segment their new users and interact with them differently, depending on which segment they fall into.
3) Pay attention to your customers’ actions.
Amazon does this incredibly well – and it’s a big one. They send email reminders when you’ve left items in your shopping cart. They alert you to sales when certain items you’ve spent time viewing drop in price. And they do it in a conversational, non-salesy way that makes you feel like they hand-picked the info for you.
4) Consider your users’ preferred medium.
Figure out which channels your customers use to communicate the most. If Twitter’s their thing, reach out to them on Twitter. For Facebook, send them a Facebook message. For email…you get the point. Doing so can help ensure a response…and also make them feel more at ease.
(And when you consider the preferred medium, you also need to consider the context and tone of that particular medium. Remember that, for instance, people don’t go on Facebook to be “sold to” – so adjust your messaging accordingly.)
5) Choose tools that offer open APIs.
The key to adding context to your customer interactions is unifying all your data points. If you use a tool that holds your data hostage, then you’re limiting your ability to get to know your customers and create a great customer experience.
That’s why we think it makes sense to look for business tools that offer APIs for ease of integration (it’s also why we built our address book app on top of our own API). You want to be using apps that talk to each other regularly, so that you can make connections, eliminate the data silos, and see each customer as a person, rather than a metric.
Obviously, I’m not against automation – on the contrary, marketing automation can be an effective and powerful tool, when used correctly. I’m simply opposed to treating every user as if they’re the same person.
Unlike dumb mass email personalization that still persists from five years ago, the above recommendations are based on what your visitor has done in the world, and what they actually did when on your site. When a user gets a message from you based on their own actions (or their interests) – provided it’s done well – they feel listened to. And as a customer, feeling like a company is listening to you is huge. (Which is one reason for the meteoric rise of social media as a customer support tool).
Ultimately, there are a million ways to add context to your marketing efforts. The key is that phoning it in via dumb automation and personalization is no longer okay. That sort of thing has become the bare minimum. If you really want to knock it out of the park, go the extra mile and add context to each user interaction – even the automated ones. You’ll see huge results.