Omni-Channel Optimization: There is no Golden Path


Share on LinkedIn

My daughters have recently begun to stage weekend “film festivals” where they will run four, five or even six movies over the course of two or three days. It began with a Best-of-Pixar marathon that included Toy Story, Finding Nemo, Monster Inc., Incredibles, Ratatouille, and Up. “Success” lead to a Miyazaki festival (Kiki, Cat Returns, Castle in the Sky, Totoro, Howl) and, last weekend, a best of Dreamworks that included Monster v. Aliens, Shrek, Madagascar and Kung Fu Panda. Perhaps that’s thinner gruel, but I love Kung Fu Panda and my topic today brought to mind Kung Fu Panda’s “Secret Ingredient Soup” and the completely blank sheet that is the legendary “Dragon Scroll”:

Mr. Ping: The secret ingredient is… nothing!

Po: Huh?

Mr. Ping: You heard me. Nothing! There is no secret ingredient.

Po: Wait, wait… it’s just plain old noodle soup? You don’t add some kind of special sauce or something?

Mr. Ping: Don’t have to. To make something special you just have to believe it’s special.

[Po looks at the scroll again, and sees his reflection in it]

Po: There is no secret ingredient…


So what’s this got to do with omni-channel analytics? Well, way back when we first started in web analytics, we used to believe that if we could only see the paths that visitors took on a website, we could really optimize things. We’d know the “golden path” that got users to convert and we could steer people to those pages.

So you know what we did? We wrote our own pathing tool! Yes, before SiteCatalyst could do this or GA even existed, we’d written a C++ program that took weblogs and created a full path analysis. I can still remember the first time we cranked it up and ran the logs of one of our few clients (a large brokerage) through the pathing tool for the first time.

It was just a month’s worth of data but it ran. And ran. And ran. Hours and hours. We killed it. Added diagnostics. Tried again. It ran, and ran, and ran. And what those diagnostics were telling us was startling. The number of different paths was astoundingly large – creating a massive in-memory hash table that eventually slowed our system to a crawl. Conversion paths were mostly long and complex. When they weren’t, they were dull and uninteresting (Homepage, Click on Open and Account, Navigate through Account Open Forms). The most common paths were almost all trivial. The more we looked and analyzed, the more clear it became. Just as Mr. Ping’s “Secret Ingredient Noodle Soup” had no secret ingredient, there was no such thing as a golden path to conversion.

If you’re at all experienced in Digital Analytics, you’ve no doubt made this discovery for yourself. There are now great visual tools for path analysis. Tools that work and are much, much better than our home-brewed path-building code. But path analysis still isn’t much use. There are too many paths, there is too much pre-built structure in websites, and customers are too different to make charting paths useful. Deeper techniques are required for interesting analysis.

So website analysts have mostly moved on from “golden path” questions. I still sometimes get asked “what’s the best path for conversion” but the question isn’t that common and usually denotes someone who’s completely unschooled in digital measurement.

But if folks in web analytics have moved past the mistaken belief in a “golden path”, the same cannot be said for people who are starting to think about and try to optimize multi-channel conversion.

Everywhere I go, my clients are building large-scale analytics warehouses with the express intent of understanding the customer journey. Don’t get me wrong – that’s a good thing. But all too often underlying that investment is the expectation that the investment will yield a single simple answer – the golden, omni-channel, path.

It won’t. I am confident that multi-channel analytics will prove to have exactly the same problem as in-channel digital analytics. Customers don’t cooperate. They don’t use the channels the way or in the order we expect. They bounce around in ways that hard to pre-calculate and that don’t always reflect a logical ordering. The better you get at measuring those choices, the more spread there will appear to be in what the optimal path really is. Trying to force your customers into a golden path won’t work, and if you try too hard, you’ll make each touchpoint worse for a significant number of customers.

Not only isn’t there a golden path, trying to create one will make your omni-channel experience worse for many – probably most – customers. You have to think of omni-channel as a network of nodes (touchpoints) with paths running in all directions. Your job isn’t to optimize for the best path through the network, it’s to optimize each node to help the visitor based on the path they happen to be on.

If the expectation is that your analytics warehouse will reveal a golden path, then the actual results are likely to be disappointing and unused or, even worse, used and failed.

So what’s the point of multi-channel analytics if there’s no “golden path”. You need to be able to optimize each customer touchpoint to be as effective as possible for that customer at that moment. How do you do that? Not by deciding in an abstract fashion what that touchpoint is used for by the average user. That’s a terrible answer in today’s world. Instead, you do it by tracking each part of the customer journey, understanding where the customer is in their journey when they arrive at a touchpoint, and then tailoring that touchpoint to be maximally effective for that person at that moment.

Omni-channel analytics is about journey analysis (as you probably thought) and touchpoint personalization (as you may not have realized if you’ve been thinking along the lines of a golden path). These two analysis types are deeply related. The key variables you need to understand the optimal configuration for a touchpoint are who the user is and where they are in their journey.  So the results of the first analysis (where a customer is in their journey) are the inputs to the 2nd analysis (what’s the optimal experience at this moment). Structure and approach your analysis that way, and you’ll deliver interesting and important omni-channel measurement.

Though there was no secret ingredient, Mr. Ping’s Secret Ingredient Noodle Soup was really good soup. And omni-channel analytics is still really useful even though there is no golden path.

The key is simple.

  • Know where they’ve been
  • Know what that means for where they are now
  • Make every touchpoint optimal for that customer

Because sometimes, knowing there is no secret ingredient is the secret ingredient. And don’t worry…there is no charge for awesomeness!

Republished with author's permission from original post.

Gary Angel
Gary is the CEO of Digital Mortar. DM is the leading platform for in-store customer journey analytics. It provides near real-time reporting and analysis of how stores performed including full in-store funnel analysis, segmented customer journey analysis, staff evaluation and optimization, and compliance reporting. Prior to founding Digital Mortar, Gary led Ernst & Young's Digital Analytics practice. His previous company, Semphonic, was acquired by EY in 2013.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here