eCommerce Café: Andy Taylor

At SMX Advanced, we spoke with Merkle’s Assistant Director of Research Andy Taylor, who also spoke on the Mad Scientist panel with our founder Andreas. Catch the full interview behind the scenes, where Andy weighs in on the future of Google’s roadmap, KPIs to prioritize, and the reality behind agencies in the SEM industry.


Direct Link

For more, follow Andy on Twitter at @PronouncedAhndy.

Full Transcript

ANDREAS: Andy, great to have you here. Could you explain to us a little bit what do you do at Merkle?

ANDY: Sure. I’m the Associate Director of Research, and so I’m responsible for just analyzing how different data trends are taking hold on the digital marketing channels that we manage. A lot of that is in paid search, but also cross-display advertising, comparison shopping engines, pretty much any major digital marketing channel, I tend to take a broad look at how our advertisers are performing.

ANDREAS: You saw that Google recently shared its product road map on Google Marketing Next, and there are many things like custom in-market audiences, data-driven attribution in general, I think, if you take a look at the market, you see that everything goes into the AI direction. AI is taking over. What do you think are the implications on our industry, on the long run?

ANDY:  Over the past few years, if you look back at Google’s past announcements, they’ve kind of always been geared towards these automated options that Google comes out with either expanding the use of the dynamic search ads or expanding their automated bidding tools and so on. I think this is just, kind of, a natural progression of that push, that Google wants to make to go ahead and create tools that can allow advertisers to turn exclusively to Google for their campaign management, and try to use Google’s information as best as they can in terms of managing campaigns. In terms of what that means for the industry, certainly, it’s long been a concern that’s going to, at some point, disintermediate agencies and high-level campaign managers. But I think, really, what will end up happening is, it will just disintermediate the worst tools out there. And so, if you have very high-value bidding platforms or campaign management solutions, then you’re still going to have a place in the paid search world of the future, but it’s really those, kind of, lower level campaign management tools that are going to go away as Google’s tools become even better and they start incorporating things like AI and machine learning into their automated tools.

ANDREAS: From a people perspective, employee perspective, what do you think will happen? So, we have less people in the future working in our industry…?

ANDY:   Right now, it seems that way. I hesitate to say for sure that, you know, there won’t be new opportunities that will present themselves, that will require more people over time. And so, certainly, every industry is moving towards less people in it, because a lot of jobs can be automated. And so, certainly, from the looks of it, we should expect a leaner industry in the future, but there can certainly be new opportunities that require a lot of heavy lifting from humans, and so that could totally reverse that trend towards fewer people.

ANDREAS: How would you position the agencies on the one hand or tool providers on the other end in this environment?

ANDY:  I think, on the agency side, you need top-level people who are doing great work, that’s a huge value-add from the agency side, it’s just having somebody constantly looking into account performance, so really, training your people up and having really smart campaign management day-to-day is going to be key to proving your value as an agency. From a tool provider standpoint, you really just have to create the best tool as possible out there. I mean, that’s the best way to guarantee yourself a place in the search industry world and, so, just try to make whatever it is you’re focusing on is as best as it can be.

ANDREAS: Merkle’s working with several large ecommerce retailers. What KPIs would you recommend a retailer to use, and how does this compare to the reality?

ANDY:   So the long-standing KPI of choice is return on ad spend. And so, that does lack a little bit of nuance in terms of things that you can bake in at this point. Things like margin, to get more profitability, as opposed to just sales revenue. Things like customer acquisition and incremental value of your advertising, and so there’s a number of ways you can bake that in. You can shoot for cost per new customer acquisition, you can try to bake in lifetime value, in which you assign greater value to an order from a newer customer because you believe that ad click is going to produce clear revenue in the long run, than what it produced directly. And so, in that way, try to make your targets as targeted to your overall business goals as possible, which, historically, have been profitability and gaining new customers.

ANDREAS: If you run ad campaigns today, you just take plain numbers and you assume that this is the true impact of these campaigns, but we do not really know what the real incrementality of these ad campaigns is. In other words, what would happen if you just shut them down. It’s even more difficult when it comes to retargeting. So, if you target website visitors, part of them will have, of course, made that decision already, and they just take it as a shortcut at buying. Do we live in a world of illusion where we don’t want to confess to ourselves that we lack a lot of knowledge and that we, in the end, make decisions out of gut feeling?

ANDY: I’d say that’s true. There is a lot of false confidence in paid search, because you can track so many different metrics, and you feel like you have the full picture, but at the end of the day, like you said, it’s very difficult to, without doing extensive testing, understand what the incremental impact of your ads is. And so, obviously, there’s some cost to setting those types of tests up. It takes up your time and analyzing those tests, things like that keep some advertisers from even going down that path, and so they just take the performance metrics at face value. But I think advertisers are steadily getting smarter and smarter about trying o figure that out. Trying to test when is that actually producing incremental value for them, versus when it’s just being attributed with the driving value that really came from other places.

ANDREAS: I sometimes have the impression that it’s kind of an alliance of different people. It’s the agency, who loves to spend more money, drive a bigger return, so they will never say, ‘My retargeting isn’t effective.’ Then, it’s the guys managing marketing. They see the great numbers coming in, and they don’t want to go to their boss and tell them, ‘Look, I’m wasting your money for years but I’m not telling you. And it’s the boss who has sold these results to the CEO, and he also doesn’t want to get back. So, it’s kind of – everyone somehow knows this is not all true, but no one really wants to lift the carpet.

ANDY:  Sure. And so, yeah, I think there are a lot of incentives to be disingenuous about how confident you are in performance results, certainly. I think people, like you list of, there’s a lot of stakeholders that benefit from the numbers looking the way they look, and so to say, ‘We need to reduce spend because we’re not actually driving value with this ad, with this audience.’ – certainly it’s not going to be a popular decision across all of the stakeholders in a particular situation. But I like to think that, at the end of the day, most people are fairly upstanding, at least, about trying to be ethically sound in how they present numbers. I mean, certainly there are some cases where that’s not true and you see it published in the industry, very ridiculous case studies and disingenuous numbers. But I think, for the most part, I think a lot of people are trying to get better about being better at measuring things, but at the same time I think the knowledge to do that well is still lacking at this point. We’re just gradually getting there.

ANDREAS: I think one of the best examples, and you still probably see this being around, is blended KPIs on brand and non-brand, where some high-level executives are satisfied with whatever, ROEs of 5, but they have no clue that it’s just driven by a brand which would come in anyways. Otherwise they would have shut everything down for ages already. But no one wants to talk about it.

ANDY:  Obviously, yeah, that’s certainly – if you’re blending brand and non-brand, that’s one of those core…

ANDREAS: It’s kind of disappearing, because marketing is getting a little bit smarter, but we still see this sometimes.

ANDY:   Yeah. I agree. And there are definitely some basic things that need to be going down in terms of splitting brand vs non-brand, at least trying to segment audience vs non-audience, even if you’re not going to turn ads off to a particular audience, at least seeing how that audience performs relative to everyone else and trying to adjust your targets, things like that. Yeah, there’s certainly some boxes that need to be checked in order for you to count as being at least somewhat ethically sound in how you’re presenting your performance, and putting those numbers up the food chain.

ANDREAS: When you launch campaigns for new clients, when you are close to reveal, what are the low hanging fruits, what are the key deficiencies you see in the campaigns when you start?

ANDY:  Sure. I mean, it totally depends on the account, I would say. If you’re talking an e-commerce account, certainly Google Shopping is now a huge share of traffic, so that would be the first campaigns we’d take a look at. I think you guys see, like, 80% of all traffic coming from shopping, for our data, I think we see about 76% non-brand coming from Google Shopping now, so that’s going to be the first place we look for in an e-commerce account, and try to identify any products that need to be split out as their targets, see if there’s possibilities to break out several campaigns in a kind of query map, to get any brand traffic to a shopping campaign and then get any manufacturer-specific traffic to a different shopping campaign – and so to try and segment the best you can, see if you can get even more value out of those. On the text ad side of things, certainly just making sure they have all their key words built out. Making sure that keyword negatives are up to date, and you’re illuminating as much poor matching traffic as possible, whether that traffic is coming through close variants or broad match, or what have you. And then also just evaluating the bidding and how efficient you are across different segments of your account. So, has the desktop return on ads been the same as phone and tablet, and are all of the different categories within your product selection all performing at the same return of ad spend, or the same profitability, or whatever KPI it is that that advertiser is shooting for.

ANDREAS: Andy, thanks a lot for your time.


Mallory Chen

Marketing Strategist at crealytics. Based in Berlin!