Video Deep Dive: Final attribution...what prompted the customer to buy? - Local University

Video Deep Dive: Final attribution…what prompted the customer to buy?


Mike Blumenthal, Mary Bowling & Ed Reese take a Deep Dive into the topic of attribution.

This is our Deep Dive Into Local from August 8th, 2017. In our Deep Dive series, we take a closer look at one thing in local that caught our attention and deserves a longer discussion.

If you have a special topic you would like us to discuss for the Deep Dive in Local, please reach out to us. If you would like to be on one or the other of our segments, reach out and send us the topic and your availability.

If you are interested in sponsoring this weekly show also please let us know.

Our weekly discussions are also now available as a podcast as well. SUBSCRIBE TO THE PODCAST HERE.

Mike: Hi, welcome to Deep Dive Into Local. This week we have a special guest, Ed Reese. And we're going to be discussing attribution, final attribution, and hopefully how it relates to local -- but in general, the difficulties that businesses have in knowing where, when, and how a client made both the journey and the decision to make that final purchase. So I'll turn it over to you, Ed. Kick it off.

Ed: Yes, so this is a tough deal and it's been one of the things that we've really been trying to do a better job of, especially the last couple of months, because it keeps coming up over and over again. So, what I see, and we talked about it prior to our call today, it sounds like we're all having problems with it.

What's really causing the sale? And we're still living in a last-click world, and there are a lot of reasons for that. It's what people know, it's easier to track, and in a lot of cases it's how agencies and people in the past, they've just always done it that way.

Mike: Well, there's also something very visceral and valuable about a phone call that you know you triggered. Both the business hears it and you know that you were the cause. And there's something to be said for that degree of action, right? I mean, when you get further away from the phone call, some of this stuff gets pretty loosey-goosey, right? I mean, the phone call's almost a cha-ching, not quite but it's close, right?

Ed: It's super close to the cha-ching but let's look at an example. So let's say I'm buying a new truck, okay? I've got a '92 Ford, it could fall apart any day now, right?

Mike: Okay, that part I believe.

Ed: Yes. It's only 25 years old. Let's say it doesn't make it to 26, okay? So I would need to have some level of awareness of the type of truck that I would want. I'm going to find that from advertising, referrals, this whole messy system. Like it is messy, it is convoluted. At some point, I'm probably going to pick up the phone or walk onto a truck lot and figure out where to go. There's all sorts of influencers before that last deal.

And it's just really tough to do. And I think part of the problem is that because it's so tough to educate, I think we're letting people be essentially blind to how the world works. You know, there's.... go ahead.

Mike: Well, I was going to say, you know, in the traditional funnel of awareness, consideration, and conversion, right, awareness -- most small businesses, I having been one for many years, allocate the awareness issue to the manufacturer of the product, right?

I mean, when you're selling Apple, really you don't wanna be at the top of that funnel necessarily. You might play there a little bit but it's hard, expensive, it's complicated, you don't have the resources for the writing or the photography. So you typically, at least in the local market, are waiting for what falls your way after they've done the awareness and perhaps even the consideration part, right?

Ed: But it depends on the industry.

Mike: It does depend on the industry.

Ed: Let's look at that example, right? So like in the car industry, they call those the doughnut ads. You have the doughnut, which is, "Hey, here's this professionally produced ad of the new Ford F150," or whatever. And the last five seconds you get to tag it with, "Big Ed's Trucks in Spokane, Washington at 27 Main," right?

And then you just put in your thing at the end and you take advantage of all this awesome, glossy photography. But most businesses have to do both. Like most of my clients, I've got a couple that are national like the truck example, but most of mine are regional. So if you're a regional to local company, I mean, let's say you're a hotel or you're anything that doesn't have a national presence, well, your money's even more guarded and, you know, valuable. How are you going to split that up if you have to make people aware that your little niche hotel exists and that it's nice and that people should consider it and it's a good deal and then the phone call, right?

So, you know, we're booking a trip right now. Our vacation is next week so we're looking between here and the Washington/Oregon coast. It's fascinating how I'm doing that search. And part of it is awareness and part of it's just specifics on price. So how do know, like in my example, I'm reading articles about cool hotels in Washington and Oregon in little niche places to stay. I'm also literally going on Google Maps and I'm just typing hotels and I'm looking at the road and I'm clicking on the places that have a price that might be a good fit.

So there's local search, there's also the hotels that I'm finding from a PR and from, you know, some sort of influence, whether it's in the news or a newspaper or a blog or whatever. What's really causing me to book the hotel I did in Yakima, Washington on the way back?

Mike: Right, but I doubt it's a doughnut ad, though, right? In other words, one of the things we're seeing is just the lack of interest in responding to ads and the difficulty of getting people to respond to ads. That most people, at least in the research I'm doing, they don't trust ads and they would rather read a review. They would rather read some sort of newspaper article or some sort of post or educational thing, particularly crowdsourced stuff, right?

So Google becomes particularly relevant in consideration in terms of where is this place and is it physically related and what are people saying about this place? So Google's playing a huge role there, maybe even larger than Facebook. Although, some of that's happening on Facebook and Trip Advisor and Yelp as well, right?

Ed: But yea, from like...oh, go ahead, Mary. Go ahead.

Mary: So, I wanna bring a different perspective to this. Because a lot of small businesses, they are not willing to invest and they're not even interested in knowing where every single lead came from and what caused it. If their phone is ringing, they're happy. If their phone is not ringing, they're not happy. And they really don't care where that came from. And most of them have a set budget that they can devote and that's what they're going to devote and the marketer has to decide how to best spend that money to get that phone ringing.

Mike: Right, but if they're not sharing with you or you don't have any control over what their YP spend is, or what their billboard spend is, or what their TV spend is, you know, then you're in trouble...or radio ad spend...because all of a sudden 80% of their budget's going and it may or may not be generating calls.

Mary: Very true. But I had a...let me just tell you, I mean, this is a true story. I had a very prominent divorce lawyer in a southern city who...I don't know why he came to me because he already ranked very well for "divorce (city)" in everything he did. And he told me he didn't care and he wasn't going to devote any money to setting up analytics so that he can tell anything more than he could already tell. And I don't think this is that uncommon. He was just very adamant about it. I think a lot of them feel that way. It's like, "Why should I pay you thousands of dollars to track something that makes you look good?"

Ed: So I agree. I think that is most small business owners. What's fascinating is seeing what happens when that isn't the case. So we have a client, they do invoice factoring. It's like large, quick loans for manufacturing. If you need like $5 million a month to fill the order that you gotta do or whatever.

So the cost per click ranges from $60 to $80 per click in this space. So it's a very competitive space. And we're tracking cost per lead. So their cost per lead usually is around $1,800 per lead, the cost per lead, and it takes roughly 20 leads to get through to get one good one, right? So it's a very expensive business.

What we decided to do was do pre-roll ads on YouTube for awareness and education, and we were getting those at around 11 cents per view. And we noticed over time, when we isolated that, that brought their cost per lead down to about $14.50 per lead. So whatever that is, $300 per lead less expensive.

And we didn't have the UTM proof because it was just too mushy and all that. But we saw that when we made that one change we saw a $350 drop in cost per lead because of the awareness and education. So I think you're right. As a marketer, to be able to have that gear to shuffle and change as you see fit and then track the success, awesome. But it's tough because most clients don't know and most marketers aren't going there either. So it's kind of a, "What do you do in this space?" kinda question.

Mike: Right. So, you know, this question came up with Barbara Oliver, which is my pet client, so I don't charge her so she doesn't really cover the actual cost. But I was sure to implement UTM codes for her. And I implemented, well, events before it was part of the tag manager but tag managers so that we could track some of this stuff. And then, I convinced her, after she got to three employees, that one quarter a year they should track all sources at the counter of all new business, and they agreed. They agreed that post-sale they would interview briefly every client.

So this was a way for them to understand in a broad sense their whole marketing picture without me being intervened. Well, without me having to say, "Oh, here's your stats." Well, like I was showing 350 conversions. She got 140 new sales of which 80 were from digital, clearly.

And so that practice on her end, Mary, and I think this is what you recommend. You say, "Look, you're not trying to make me look better, you're trying to help you spend your money wisely." And if you're getting a good response on radio or YP, whatever, and what she learned was YPs, they did get some response. It was mostly over-65-year-old bitties who wanted a $10 repair, right? Well, okay, that wasn't what they were looking for.

And the radio, which was $10 grand, very expensive but it generated $30 grand in direct business because it was mostly men buying expensive gifts for their wives at the last minute. So they were able to ascertain that. And then they came back to me and said, "You know, we're spending $300, $400 a month on social/SEO, is it worth it?"

And that's a digital attribution question, and I went back and, some of it we couldn't track pre and post because they built in a website in the middle of that. But it was pretty clear, when we looked, that there was some real significant benefits, again, a little further up the funnel.

And it's hard in a really small business like that to really get down to where did that actual...I mean, it's hard to know across platform because it's hard to even know attribution at all, but we were able to get some attribution.

Mary: And attribution has become so complex that it's really hard for the client to even wrap their head around it anymore.

Mike: And David has often said that, you know, most small businesses like you say, Mary, aren't willing to pay for the information. My sense is, if you can deliver low funnel data in a really clear, concise way that they may not be willing to pay for it, but they would be willing to read it, which I think is part of the issue, right?

Ed: Yes.

Mike: And if you can get them to read it and it's meaningful, like how many calls came through these channels, right? How many driving directions came through these channels? And is it worth shifting between them, right? So I think there is, at least at the very small level, not bigger levels, they have bigger questions. But, very few people in the local space, outside of car dealers who have been tracking cost per lead for a long time, very few people are actually doing that math, right?

Ed: Yep. Yes, so I guess we conclude with, "Attribution is hard." It's hard.

Mary: Attribution is hard.

Mike: An example of attribution being hard, right, I decided to use campaign codes for them, particularly vis-a-vis Google where clearly 85% of their business was coming out of Google. We wanted to ascertain whether it was the service menu on Google or the regular website on Google local or whether it was organic. And using one of the data distribution services, they managed to periodically overwrite the UTM code at Google. So it was hard. Because even when you do do it, there's other moving parts in the ecosystem that are going to goof you up, right?

Why shouldn't Moz or Neptune or Bright Local or White Spark support -- I mean White Spark probably does, but you'd have to make a special ask, you know -- to use UTM code so at least you know last point of attribution source, right? But even then, when I was trying to do the best for them, it got overwritten.

So the question I have is, if you can't even get the basics of tag manager and UTM codes implemented, how do you move on to cross-platform attribution? What's the next...let's assume I could get it stabilized and get a good view, how do I then look further up the funnel to understand which activities up the funnel were really meaningful? I'm asking. That's a question, not rhetorical.

Ed: No, no, no, but the question, that goes back to what we talked on the general video earlier -- it has to be defined. So if we don't have the right assumptions on what's actually making that impact, it's not going to matter anyway. So what we see when things are set up from a funnel standpoint of what is the defined awareness, well, those are all based on assumptions. So, how do you test those assumptions? And in my opinion, it's less about the tool and more about taking the time to go through and do isolated experiments to see if you did anything.

Like you asking Barbara Oliver to ask customers once a quarter or one quarter a year or whatever it is, us doing that isolated experiment for YouTube pre-roll for the financial client. And that takes a lot of time because then you're looking at every step of that funnel doing some sort of independent test, isolated test to then get the baseline, to then get the funnel, so that then you could set up the stuff and the things to try to measure it, right? So it's a lot of work.

Mike: Right, it is a lot of work. And it's hard to understand. Like in Mary's case, I think a lot of people are not willing to pay for it. I think there's an opportunity there. I mean, again, I don't know. I think Google is missing an opportunity. For example, why doesn't their analytics read insights and have a dashboard report that's clear and simple within analytics that reflects insights, right?

I mean, how can you explain that huge missing piece there? Because Google Insights tracks very low, final things are valuable to businesses: calls, driving directions, photo views,'s hard to understand even with photo views, though. What is the value of a photo view on Google? How do you track....? And again, not a rhetorical question, how do you convert the value of a photo view on Google to a meaningful attribution at the store level, right?

Ed: See, that would depend because you could do that in certain industries. So jewelry, that could make sense. Hotel, that could make sense.

Mike: Right. Oh, I get it, it makes sense. The question is, how do you know though anything? If Google says, "Oh, your pictures were viewed 10,000 times last month and it was eight times more than the nearest competitor," what does that mean? And how do you then ascribe some value to it, right? I mean, maybe you say, "Okay, for every 5,000 photo views, we're seeing 500 phone calls at Google or something." Maybe that's how you do it.

Mary: But they may not have any relationship to each other.

Mike: Well, a funnel is, like Ed said, a funnel is based on assumptions. And if like in the jewelry industry, one assumes that all those views are valuable, the question is, what is the value and what happens next? And Google doesn't provide us with any insight as to the behaviors -- really what we need to know is how many people clicked on the photos and then went back and did something else that was significant, right?

Mary: Right.

Ed: In closing, I can say that we did a blog post like three or four years ago on networks and like a circle of trust and kind of explained the circle of trust comment. And Tine did the drawing and her circle of trust basically it looked kinda like this. It was basically just a circle with a dot in it. It kinda looked like a breast a little bit. It wasn't intentional, it just kinda looked that way. She drew it in like four seconds. If you do an image search for circle of trust, it's the number one image on the internet and has been for years. I'm pretty sure the circle of trust image has not helped our agency, right, even to the tens of thousands of views that people see our circle of trust.

Mike: You think rather than reflecting Tine's incredible skills, it reflects sort of a Freudian slip that she wasn't even aware she was making?

Ed: Potentially, I don't know. Well, I think that pretty much sums it up for the day.

Mike: All right. That's a great way to end, I think. I don't know, we'll see, we'll see. All right, thanks for joining us for Deep Dive. We'll talk to you soon, bye bye.

8 Responses to “Video Deep Dive: Final attribution…what prompted the customer to buy?”

  1. I just finishing listening to this. First off I want to thank you all. You are among the elite in Local Search and I respect everyones opinions and views on this panel. I am a one-man shop who started out in 1986 doing design. I did my first website more than 12 years ago. I always try staying on top of the trends in my field so I can continue to add value to my customers. Local Search was a nature path I followed knowing it would be good for my customers, therefore good for me.

    Fast forward to today and I am shaking my head at your conversation. I mean you all at are the top of your game. It first sounds like you have a tough time quantifying the need to track measurable, meaningful data for the good of the customer. The customer does not want to pay for that, they just want the phone to ring. How true that is!

    By the end of this I got the impression that even you are admitting that knowing what you know illuminates how much is fuzzy math for reasons out of the control of even the most savvey in our field.

    My takeaway from this session is that the customer just wants results (i.e. phone ringing, people purchasing because as a result of something we can attribute our hard work to). They do not care how that is achieved. Show Me The Money is and always be the bottomline.

  2. I think you nailed it… I think metrics are important but the tools are inacurate and client’s eyes glaze over….

  3. Just read this. My general reaction is the same as the one respondent, Chris: Actual attribution is tough, and in many cases the client glazes over this.

    And then you confirmed. Oh my oh my. and this deserves a Laughing Out Loud. Much deeper than an LOL.

    Mike: As you know the smb in my sig and also some others do the kind of attribution research you did with Barbara Oliver. We do it ALL THE TIME. We have thousands of responses. They comprise a significant sample of all sales.

    It is deep. Its much more telling than any web stats. We use it to determine what works and what doesn’t work. Then we do put more effort/$ into the things that work and drop the things that don’t work. Then we try new things. If they work great. We double down on effort. If they don’t work we drop them.

    I have a question or series of questions:

    When you wrote 350 “conversions”. What does conversion mean? Were those the total number of customers that showed up and you interviewed? Was it something else?

    I assume by counting the 140 sales you were also able to ascertain that among them were the little purchases that involved an inexpensive repair and where the purchase came from. I also assume you could track the actual $$ purchases to the people (husbands) that learned of the store through the radio. And so on.

    Are my questions on the right track or not? I understand 140 sales, but what does 350 conversions mean???

    If there were 350 customers in the store over the period you were interviewing them, of the 210 that didn’t purchase that day…do you know if they came back to purchase???

    Not everyone buys on day 1. Lord knows when I used to shop for jewelry with my ex we shopped (till I dropped–she could go on for seemingly ever). but we always finally bought. Sometimes it took multiple trips to multiple stores before one of those stores got our $$$ and we got an item(s).

    First visits might turn into later sales. Was your research accounting for this???

  4. 350 conversions refers to the digital KPIs that we identified as important to Barbara that we were able to assemble from across our tracking:
    Driving Directions
    Click to Call
    Contact Form
    FB Messenger message

    140 Sales was just NEW customers. This number didn’t include

    So some of the 350 were old customers, some were lookers not buyers and some were probably no shows or dupes….that we tracked as far as we could online.

    Of the 140, somewhere along the line of 85 said they found us online…

  5. Tx.

    Of the 140 new sales, some of them could have first contacted Barbara Oliver before the period when you were accounting for KPI’s, I assume. That is often our case.

    Regardless thanks for the data. I would hope she reacted to this. A majority of the sales were attributable to the web, and it appears you confirmed that some of the other sources, while they might have generated sales, might also have generated tiny sales with little income.

    Nice work. When it comes to attribution as you described above, its tough to tie it to sales. Your effort with Barbara Oliver did so.

    I think that is quite worthwhile.

  6. @Dave
    Yes, we made progress. Unfortunately Barbara lost the person that was doing all the marketing stuff and we are starting at ground zero. But at least what we learned last year has been implemented this year.

  7. Good. We learned that as we “learned more” from analyses similar to what you did, and we made changes in spend and effort that reflected what worked vs what didn’t, it improved our performance…and by that I mean sales.

    Here is a “technical question” related to how you did your analysis, though from your description this particular type of contact wouldn’t have occurred with regard to Barbara Oliver.


    Of the smb’s we operate, one has great signage, one mediocre, one a little, and others with none. The one I report on the most has mediocre signage. Its HIGHLY Visible, but for its market territory its visibility is relatively minor.

    Still it works. After the 2 largest sources of leads; Search and WOM/Referrals its the 3rd most effective source. Small relative to the big 2 but bigger than anything else.

    The SMB with great signage gets a lot of leads that way. Very significant percentage. Sales develop from that signage.

    Here is the “technical question”. Its theoretical.

    If a customer saw signage of your business, had looked it up on desktop than called the business number via a landline (presumably having seen it in the KP). Would you catch the KPI???

    I presume G My business stats wouldn’t capture any data. The business would get a phone call. If you were trying to capture marketing data off of phone calls, you could of course get a response.

    While you mentioned that Barbara Oliver doesn’t have visibility, did your methodology allow you to capture and characterize a lead such as that one?

    From our perspective we focus separately from web data on leads. We ignore a signal such as “driving directions”. OTOH, we have a “non web” “equivalent or similar signal” to driving directions.

    When leads come via a form, we know it was a web visit. We don’t know how or why the web visit occurred.

    On the phone we do a middling/ or half effective effort at trying to find the real source of the call: ie search, referral, WOM, anything else. Our ability to get this data is only middling successful. When we are on the phone most of our conversation is going to involve the service/questions from the lead, and an effort to sell or another intermediate step.

    Hence in our case the best data we get is the lucky and unusual opportunity to interview buyers while they are using our service.

    Anyway, back to the question: How would you capture that land line phone call “kpi” if at all?

  8. I think signage with a great big number on it is one of the vastly underutlized tactics in today’s world. Every time I walk by a new local sign and it doesn’t have a number on it, I think how many people drove by and weren’t given the chance to dial their mobiles.

    As for tracking those calls and getting more effective measurement from your phone calls is to implement a disciplined call tracking system… focused on local that uses only 2 (or maybe 3 numbers) for public consumption.
    1- Is your local number that gets ported to the call tracking system and is kept for local tracking..
    2- An online number for citations that gets filed as the primary number for all on line citations while the actual landline gets listed as a secondary number every place.

    Thus any calls coming in on the new line would be from your online efforts and any new calls coming in on your local number would be from your local efforts. Not perfect but at least a stab at it.

    3- If you wanted more local granularity then get a third number for things like your sign, billboards etc where it can be used to track local sources that can’t leak online.

    Your thoughts? What would you or have you done?

Leave a Reply