Bummer! This is just a preview. You need to be signed in with a Pro account to view the entire video.
Estimating Traffic Based on Keyword Research30:14 with Jessica Bowman
Don't let changes in Google's tools get you down. Jessica's going to improve your spirits by showing you how to estimate your keywords' traffic with the data at your fingertips.
[Jessica Bowman] Laura was right. I do travel a lot. 0:00 For a while, I was actually homeless and literally lived out of a suitcase. 0:02 She is right about that. 0:07 I'm going to talk to you today about 0:09 estimating traffic based on keyword research. 0:11 Before we get into that, we need to kind of set the stage. 0:16 Who are you doing this for, what you need to be thinking about, 0:19 and why it's important. 0:22 I tend to think about who are the people 0:24 that we've got to address? 0:28 We've got, in every organization, a few different kinds of people. 0:30 We've got this analytical guy here on the right. 0:33 He's the grump, right? We all know that guy. 0:37 We all know the one that's going to 0:40 criticize everything and look for holes, so you've got to 0:42 make sure you think about him and provide 0:45 very thorough information. 0:47 Educate him on how you came to some of your conclusions. 0:49 The second we've got—these might be the developers, 0:53 they might be product managers— 0:57 they're going to be these guys that are calculating. 0:59 "We could do this SCO thing or we could do this thing over here 1:03 that generates millions of dollars and I know exactly what I'm going to get out of it." 1:07 You've got to think about these guys, 1:11 because they're always thinking about how can we get more done. 1:13 Maybe they're also thinking about how we can get things done cheaper. 1:16 SCO is not usually the cheap option. 1:19 The last are the executives, right? 1:22 What do the executives need to think about? 1:25 They're always trying to get more done with less cost. 1:27 The best ROIs, so it's really important for you guys 1:33 as SCOs to really think about, 1:35 "Well, how do I really calculate all of this 1:37 and think about the agendas of these different people?" 1:40 The challenge that we see with a lot of 1:44 SCO keyword research is that there's this challenge 1:48 to make the leap from the keyword traffic 1:51 that comes from the tools into actual avenue 1:55 and what I'm going to get in terms of revenue and profit, 1:59 so turning it into dollars. 2:01 What we want to think about is, 2:03 "You know what? That's what management needs, 2:06 all these guys need the stuff on the right." 2:08 This presentation is going to help you by filling that gap 2:10 so that you don't get cut, because these guys here 2:14 they're focused on what can we cut. 2:21 They're often the ones that will decide whether or not 2:23 your stuff stays in scope. 2:27 Today's presentation, we're actually going to fill that gap 2:30 by determining what's that estimated traffic. 2:33 How do we actually predict 2:35 what can of traffic are we actually going to get 2:37 for a given keyword, so that we can then 2:40 take it to the next level and say, "Okay, if I get 2:43 this kind of traffic, now I can do revenue and profit." 2:45 We're going to focus on the estimating of traffic. 2:48 We've got a lot of hurdles to cross today. 2:52 In today's world, keywords have changed— 2:55 I'm sorry, the search results have changed quite a bit. 2:58 They're not all equal. 3:00 Not all the search results pages are the same, 3:03 not every listing is going to be as valuable as other listings, 3:05 and so we've got like author photos, 3:09 review stars, numbers of reviews, 3:12 things like even recipes have 3:15 how long does it take to prepare. 3:17 There's a lot of data points that are actually 3:20 competing against your listing that may not have all of those elements. 3:22 Another thing we have to think about is 3:27 they're not as useful, these ranking predictions, as they used to be, 3:30 because the ranking predictions used to 3:35 be all for search results pages that are 3:38 identical, but the problem is they're not identical 3:41 for every user anymore, so we've got these personalized results, 3:44 and what's happening in Los Angeles is going to be different versus New York. 3:47 Also, my history, what I tend to click on— 3:51 all of these things start to impact. 3:54 You have to takd that stuff into account. 3:56 Everyone is looking for this magic box. 3:59 How do I do a tool? I generally think a tool can be created, 4:01 and I'm kind of surprised I haven't seen 1 yet, 4:05 on how to do this and guide someone for this. 4:08 But to be honest, it's going to feel a little bit like this, right? 4:11 Like you're looking into a glass ball. 4:14 We hear some giggles. It sounds like yeah, that's kind of what it sometimes feels like. 4:17 I'm going to walk you through a process for this. 4:20 My life before I was an SCO— 4:24 I did user experience and project management, 4:26 but I also had the role of process analyst. 4:28 I know that sound super cool to people like us, 4:31 but my job was to actually systematize things. 4:35 When I thought about how am I going to present this stuff, 4:37 I actually said, "Okay, right, this needs to be a process." 4:40 Something that you can take away— 4:43 download the slides, take it away, and execute on it. 4:45 How many of you guys are in-house SCOs? 4:49 Okay, you guys don't do this day to day, right? 4:53 It looks like maybe a good 40% at least were in-housers. 4:59 The consultants in the room are going to do this stuff 5:02 quite often with clients. 5:04 But I know when I work with in-house SCOs, 5:06 they do this maybe once or twice a year, 5:09 or maybe when they're going to launch a new section 5:11 or launch a new business line or product line— 5:13 it'll come up a little bit more frequently, 5:16 but so infrequent that it's going to be very unfamiliar territory. 5:18 This is a process that you can repeat and follow along. 5:21 We made a couple of assumptions. 5:25 I made the assumption that you already know advanced keyword research. 5:27 We are not going to talk about 5:31 how to gather keywords using tools. 5:34 I"m assuming you know how to do that. We're not going to talk about 5:36 how do I weight the demand versus competition. 5:38 I'm going to make an assumption that you've thought about your personas— 5:41 who you're going to target, what are the keywords that they want. 5:44 I'm going to assume you've mapped your keywords 5:47 to the buying cycle and to the pages that you're targeting. 5:49 All of that stuff's done. We're taking it after all of these things happen. 5:52 I'm also assuming your keyword research is complete. 5:57 I want you to think about this not as the perfect solution, 6:00 because you're going to find some holes in it, mostly likely. 6:04 No model is going to be perfect. 6:07 It's not perfect. It's 1 model you may need to adjust, 6:10 so sometimes we get into this and we actually 6:13 realize for this client and the data we're seeing, 6:16 my intuition says it's wrong. 6:18 We'll make some manual adjustments to the model, 6:20 so be thinking about that if you have to use this. 6:23 This is just kind of a subset of the things that we train in-house SCOs. 6:27 I only had 20 minutes, about, to give a presentation. 6:31 I couldn't do everything, so we kind of simplified this down 6:35 into something that you could walk away with a process to follow. 6:38 Before we begin, you need to know your site's weaknesses, 6:44 the competitiveness in your market, your current traffic levels, 6:48 and all of that's going to lead to where you think you can rank. 6:53 I am making an assumption that you can figure out 6:57 by looking at all of your data that I know where I'm going to rank, 6:59 so that I can figure out where my click-through data rate might be. 7:03 I"m also going to assume you know your keyword volume. 7:06 The overview of the process, so let's just walk through it. 7:11 The first thing we're got to do is predict 7:14 the actual search volume for any keyword. 7:16 We're getting the volume from tools, and the tools are 7:19 not always right. We know that. 7:22 We've got to figure out how do we predict 7:24 the actual volume for a keyword. 7:27 Then we need to predict what's our average ranking. 7:29 We have to actually think about what do I think I could actually rank for. 7:33 If position number 1 is Wikipedia, number 2 is Amazon, 7:36 and number 3 is the New York Times, I'm not likely to take positions 1 through 3. 7:40 That number shouldn't even be something I account for 7:44 in my calculation for predicting my traffic lift, right? 7:46 Okay, so then once I know my average ranking position, 7:51 then I'm going to predict what's my click-through rate, right? 7:55 There are a few studies that have been done on that. 7:58 You can use your Google or Bing webmaster data on this. 8:00 There are lots of things that you can do, and I'll talk about 8:05 how you can brink all of those things into the picture. 8:07 From all of these things, we can then estimate our lift. 8:09 The problem is, we have these first 3 hurdles to overcome. 8:13 As I talked about, if you do it infrequently, it's always going to be a challenge. 8:19 I'll just kind of walk you through it. 8:22 What we done is we have a spreadsheet, so this 8:24 spreadsheet walks you through the process. 8:26 This is part 1. Do I have a pointer? 8:29 No. Okay, so this is part 1 right here, 8:33 the average search volume, so that's step 1 that we're going to go through. 8:36 These 3 columns are step 2, 8:42 where you're expecting to rank, 8:44 and then step 3 is your estimated click-through rate. 8:47 Let's walk through all of these steps. 8:51 We're going to use a sample keyword, 8:53 and we're going to use the keyword home insurance calculator. 8:55 Let's go through how we predict the actual search volume. 9:00 The first thing you want to do is get your volume from the tools. 9:04 I'm going to make an assumption that you know how to go 9:08 get the volume from Google AdWords. 9:11 You need to go into the Bing webmaster tools, 9:14 because if you aren't familiar, they've made some updates, 9:16 and their keyword research tool is in the actual webmaster tools. 9:19 You probably actually, at this step, don't even need to do this, 9:23 because it should be in your spreadsheet from where you 9:26 did your keyword research to select the keywords, 9:29 but if you're cutting some corners, that's where you need to go get it. 9:31 The thing about this data is 9:35 that it's interesting—we have to think about 9:37 the search engine's traffic share. 9:40 We actually are recommending you go get 9:42 both Google and Bing data, not just 1. 9:45 I'll talk to you about why. 9:48 If you look at the market share here, we have Bing representing 9:50 15.6% and Google at 6.8%. 9:54 If we use that data, so we've got the 9:59 monthly average search volume for Google and Bing 10:01 for our targeted keyword. 10:04 We have here their search volume, 10:06 so we know this is the search volume. 10:09 Then we can say based on the search volume, 10:12 we can say they represent 66%. 10:15 This is 4,800 is what 10:19 we would estimate to be the total search volume 10:22 if we base it on Google's data. 10:26 If we base the total world search volume, or US search volume, 10:29 based on Bing data, we take the 237, 10:34 account for the fact that they're only 15% of the market share, 10:39 we would say, "You know that? 10:42 Bing would tell us that there's about 1,500 searches, right?" 10:44 There a huge disparity. 10:48 What we actually do is say you know what? Let's just adjust 10:50 by taking an average of the 2. 10:53 That how we'll account for that, so we're just 10:55 going to take an average of the 2. 10:57 In your world, your intuition, your experience may actually say 10:59 you know what? These tools over-report or under-report in our particular market. 11:03 You might not want to do an average. You may want to do something else. 11:07 But this is 1 way that you can actually just 11:09 kind of figure out what can I do 11:11 to come up with a search volume that's in between those 2 numbers. 11:14 All right, so then you're going to 11:19 adjust for searcher intent, so here's the problem with the search results. 11:21 There are multiple searcher intents that the 11:25 search engines account for in any given thing. 11:28 In this example there, we're saying install Windows. 11:31 When you think about install Windows, 11:36 some of you might be thinking about the windows in your home. 11:38 Others of you are thinking about Microsoft Windows. 11:41 Well, you know what? The search results account for that as well. 11:44 Some of them are presenting home renovations 11:49 and some of them representing software. 11:52 When we think about this, we have to account for it in our calculation, 11:55 because you know what? We're actually going to get fewer clicks as a result. 11:58 We also don't want to account for that portion 12:01 of the searches that are being conducted for software. 12:05 Let's say we sell Windows. 12:09 We're a company who maybe sells Windows. 12:12 We don't want to estimate our lift 12:15 based on the amount of people who are actually 12:19 looking for Microsoft Windows software. 12:23 Let's look at another example. 12:27 In this example, we're looking at the 12:29 something for our insurance, right? 12:35 In this example, we're looking at some of them 12:37 are transactional in nature, some of the search results. 12:39 Some of them are more informational in nature. 12:42 Well, if your search results account for informational searches, 12:46 that means the search engine thought that some people 12:49 typing that keyword are not interested in a transaction. 12:52 They're also interested in a little bit of some information. 12:55 Those tend to be the best results is what they're saying. 12:59 What we're doing is we're actually going to 13:01 use that information in the search results 13:04 to do some back of the napkin calculations. 13:07 We're going to say, okay, here's the search volume, 13:10 I'm going to guesstimate that you know what? 13:12 If the search engine listed 13:14 8 of the listings as being relevant 13:17 to my searcher intent, then 80% of those searches 13:21 are probably going to be a good match. 13:24 Here's an example. If we say 8 out of the 10 13:27 pages are transactional, let's assume that 80% 13:30 of the search volume is transactional in nature. 13:34 If 5 of the 10 pages is relevant to your searcher intent, 13:38 then you're going to assume 50% of the traffic will be what you're looking for. 13:41 What your searcher will be looking for. 13:45 That's how we kind of estimate, okay, 13:47 here's the total search volume, but this is the percent of that 13:49 total that's actually interested in what I'm potentially selling. 13:53 We want to make sure we're accounting for that. 13:57 In that spreadsheet, what you'll do is you'll add this information. 14:01 There's a column where you put the search volume, 14:05 there's a column where you put Bing's volume, 14:08 then we have things that are automatically calculated 14:11 based on that, so it will actually calculate those estimates 14:14 and the average, the average that we talked about, 14:18 it will do this automatically. 14:21 We'll adjust, so this is where we say you know what? 14:24 We think 80% of this page is going to be relevant, 14:26 80% of the search results page is relevant, 14:29 so I'm going to assume 80% of the searches must be relevant 14:31 to my searcher intent. 14:34 That's where you make the adjustment. 14:36 The second step—so the second part. 14:40 Now that I know what my potential market is, 14:43 I've estimated my traffic lift, 14:46 now we need to predict the ranking position. 14:48 How might you do that? 14:51 I mean, you really have not a lot of data to use. 14:53 This is probably a step that follows 14:57 a lot more intuition and experience, 14:59 rather than hardcore data. 15:02 This is the fuzzy part that when you're talking to 15:05 executives and stuff like that, you want to make sure 15:08 you communicate how you came to the conclusion for this one. 15:11 I tend to think about if big brands were listed, 15:15 because we're a smaller organization. 15:19 If a big brand was listed, we dropped our ranking a bit. 15:21 We're probably not going to out-rank some of these big guys. 15:25 Things that you can do is go to SEOmoz. 15:29 They got a great keyword difficulty source, 15:32 so oftentimes, we'll look at this. 15:35 We'll actually come in and look at the data— 15:37 how difficult is it to rank? 15:40 Well, let's look at the domain authority and the page authority 15:42 of the position ranking. 15:45 Then we've also got 15:47 some of these detailed pieces of information lower on the page. 15:49 We can do this and say based on that, 15:55 I can make an assumption. I can see who's ranking for that keyword. 15:57 Then I can see I think I can beat this guy, 16:01 I don't think I can beat this guy, I think I can beat this guy, I think I can beat this guy. 16:05 That how you can actually go in and figure out 16:09 here's where I think we can rank. 16:12 What you would want to do is define a high position and a low position. 16:14 This is how you're going to get your gap. 16:19 I think we can attract this much traffic to this much traffic 16:21 by understanding where you think you can rank. 16:24 Now, you then enter that in your spreadsheet, 16:27 so you actually just put a column for low and high ranking. 16:30 Now, you're ready for part 3, so now we know 16:36 our estimated traffic—I'm sorry, the estimated search volume 16:38 relevant to our searcher intent, 16:41 we now know of these keywords, here where I think we can actually rank, 16:43 so now it's time to predict our click-through rates. 16:47 The best data's your own data. 16:52 The challenge with this is that your own data is only 16:54 going to be relevant for the keywords that you actually 16:57 have decent rankings for, so it's really not a great source 16:59 for the vast majority of your keywords, 17:03 but you can go get your own data through the Google click-through data 17:05 and the Bing click-through data in the webmaster tools. 17:08 It's not going to help you with the keywords that you're still targeting, 17:12 so you're still kind of stuck—you've got to figure this out. 17:14 This is just an example in Google webmaster tools, 17:19 where you want to actually get the click-through data 17:22 or a given keyword. 17:24 The alternative is to look at 17:27 prior click-through rate studies. 17:30 I think 2 fairly recent—although 1 of them's a couple years old now— 17:32 sources are the Optify study, 17:36 and then also the Slingshot study. 17:40 You might want to choose 1 or the other that you find most relevant. 17:44 We tend to actually prefer the Slingshot data one. 17:48 It's a little more recent, and it recognizes 17:51 that not all searches are going to lead to a click. 17:54 It took a little bit more into account about what 17:56 the search results really look like today. 17:59 We take that data. 18:02 The downside of thinking about only that data 18:04 is that it's still not going to count 18:07 for the things that throw your 18:09 click-through rate off, such as 18:12 you have a lot of videos showing up 18:16 or you have a couple of leading authorities 18:18 showing up that would actually cause for a given keyword 18:21 click-through to decline. 18:24 What you want to do is go ahead and take the data— 18:26 the set of click-through rates that you chose, 18:30 whether it's Optify, Slingshot, or something else, 18:33 and you want to actually just enter that in the spreadsheet. 18:36 Then we actually account for if I want to 18:39 adjust that for a given keyword, 18:42 I can just enter it in each cell underneath. 18:44 I would say you know what? I know, based on what I can see 18:47 in the search results, if I rank there, I'm going to 18:50 have a lower click-through rate, so let's just break it down. 18:52 You can actually bring it down. If you do that, 18:55 you can then word by word account for this. 18:57 You don't need to do it for everything, right? 19:00 This is definitely not something easy to scale. 19:02 But for some keywords, your intuition, such as 19:05 certain keywords that tend to have a big brand showing up, 19:09 you just know those keywords, I need to give a little more manual attention. 19:12 But all the others, you just go with the Slingshot click-through data. 19:15 Now it's time to calculate the lift, so we've gathered a lot of information. 19:22 The spreadsheet kind of facilitates that, right? 19:27 You want to make sure you log it. You've got all the columns that you can put in there. 19:29 You log all this data, and so what you want to do actually 19:33 is all of the data in the orange— 19:35 so we're taking into account the average search volume, 19:38 we're taking into account an adjustment for 19:42 the matching of search, so it's an estimate. 19:45 What percentage of searches do I think are going 19:48 to have the same intent that I can deliver on? 19:51 Then we're looking at what do you think is your 19:55 high ranking versus your low ranking? 19:58 We're taking into account all the click-through rate data. 20:01 Based on that, the spreadsheet can automatically 20:04 calculate your estimated traffic lift. 20:07 We can say here's my range, so for home owners 20:12 insurance calculator, it can be between 87 and 316. 20:15 You can use this spreadsheet to kind of drive 20:21 what kind of metrics you want to communicate 20:24 out to the organization. 20:27 Now, I've done this quite a bit, and I was an in-house SCO, 20:32 and so I really remember what it was like to 20:35 communicate the data, and I had 20:38 some war stories, right? 20:41 People who have heard me speak, I often talk about my war stories 20:43 back when I was figuring it out. 20:46 What we want to do is think about 20:49 remembering these guys, right? 20:51 You've got to remember the guys that are really 20:53 there to pick you apart, especially if you're not a data person. 20:55 I know this event seems to have attracted a lot of 20:59 data presentations, so you guys are loving it who are data people. 21:01 But those who are not data people, this is especially important for you. 21:05 When we think about that, those guys want to cut. 21:09 You need to make sure you back up everything and really present right, 21:11 so emphasize, it's not easy to calculate this 21:15 and explain how you came to these estimates. 21:18 What you want to do is walk them through 21:21 here's the process, here's how I came to these conclusions, 21:24 and let them understand and build trust with your approach— 21:27 your model, your calculation model. 21:31 I really recommend you use a range, not a finite number. 21:33 That tends to show that there's a gap. 21:38 If you don't do things right, you're going to actually 21:40 get the lower number or potentially less. 21:42 As much as you can, use a gap. 21:45 Identify what needs to happen to achieve the metric. 21:48 This is where I think a lot of SCOs I see 21:51 when I would look at in-house programs, they kind of 21:54 miss the mark, because they're not specifying 21:56 clearly enough for everyone to understand 21:59 that if you don't do this, you're not going to get this. 22:03 They might talk about it, but it's kind of glossed over, 22:06 and it's not spelled out in crystal clear fashion, 22:09 so you want to make sure that you're really talking about 22:12 we need to launch 5 articles or log posts per month, 22:15 we need to hire a copy writer, and that copy writer 22:18 needs to produce something every single day. 22:21 Whatever it is, you need to make some assumptions 22:23 on what's going to happen to achieve that metric. 22:25 It's not just for a little bit of CYA, 22:29 but it's also for them to understand 22:31 what do I need to account for in the budget. 22:33 That's why that's going to be real important. 22:35 I also recommend you identify all risks 22:38 that could hinder the traffic goals. 22:41 You want to make sure you think about URL strategies 22:44 not being search friendly, the CMS is not going to be upgraded, 22:47 anything that can go wrong that will 22:50 cause you not to achieve your results, you want to talk about that. 22:52 This is especially important if there's something 22:55 that you're trying to get through the system, 22:58 and you know that if you get that through, you would achieve this. 23:01 You want to make sure executives truly understand this. 23:03 Then you want to make sure you're adding dollars to the presentation. 23:07 Now that you know here's what I think my traffic should be 23:09 for any given keyword, you can then easily 23:13 take that leap into revenue. 23:15 You might be able to use some pay per click data to drive 23:17 some of that decision, you might be able to use historical data 23:20 to drive some of that calculations as well. 23:23 Also, you want to say you know what? 23:28 For us to get this kind of traffic through other channels, 23:30 this is what it would cost, so then you're giving a comparison 23:33 of I know I'm asking for this, this, this, and this, 23:36 but here's what's on the table, and for us 23:38 to get that same amount of traffic, 23:41 this is what it would cost through other means. 23:43 That will actually help you with getting SU prioritized. 23:46 That's all we have. Thank you very much. 23:49 [Applause] 23:51 >> Thank you, Jessica. >> You're welcome. 23:59 >> I see hands up already. 24:01 Start right here. >> Yeah. 24:05 >> Hi. I just wanted to ask if you could 24:07 make that Excel spreadsheet available to us, 24:10 because it's a really nice page. 24:12 >> I can. It's not posted online, but 24:14 you email firstname.lastname@example.org. 24:18 That goes to my assistant, and I will tell her to give it out. 24:23 I didn't think there was a thousand people here, 24:26 so she doesn't know that. 24:28 >> Okay, thank you. >> Yeah, we can give that out to you guys. 24:31 Just email@example.com, or you could just drop a card, whatever. 24:33 >> Hi, thanks for that. That was really useful. I'll be trying that when I get back. 24:45 Do you ever factor in Long Tail onto this? 24:49 >> Yes. Sometimes we use a little bit 24:53 of a different approach for Long Tail, 24:56 but I think you could still do this, so take in 24:58 all the search volume for the Long Tail. 25:01 Sometimes we start just querying what kind of words contain 25:03 this theme or what kind of words contain that theme, 25:07 but it's not as systematic. 25:10 It's a little rough how we do it. 25:12 But I really felt like you could probably 25:15 take your Long Tail keywords that you've identified in keyword research 25:18 and throw it in here and get the same kind of results. 25:21 >> Sorry, what about like, say, you create a new page, 25:28 you're targeting a head term, 25:31 it's not just going to be that head term that come up. 25:33 There's going to be some stuff, and there's going to be stuff that you're not going to be able to 25:37 get search volume for 25:39 necessarily out of the tools that are available. 25:41 Do you add 8% for that stream Long Tail, 25:43 or do you just— >> I'm sorry, repeat the question again? 25:47 >> Say you create a new page for a new product or whatever. 25:52 It's not just going to be head terms, and it's not necessarily 25:55 just going to be terms that have search volume 25:58 that you can find through the tools that are available. 26:01 Do you just say 8% 26:04 or however many are going to be really random Long Tail stuff, 26:08 or do you just not say anything? 26:11 >> I think that that's a little less about the model 26:15 and more about thinking about what kind of keywords 26:17 could it actually rank for. 26:20 I would be taking a look at first, what do I think 26:22 the combinations could be based on the content of the page. 26:25 It sounds like you're trying to predict 26:29 before we know what's on the page, 26:31 so you're early in if we think about 26:33 the time it takes from idea 26:35 in prioritization to launch, you're up here 26:38 trying to do the calculations, but the content is written here. 26:41 The challenge with that is you need to know what's going in the content 26:44 to truly predict this, unless you, 26:48 as the SCO, control the content. 26:50 If that's the case, if you don't control the content, 26:53 I will wait until content is done, 26:56 but you might be able to say my goal is it ranks 26:59 for these 20 words or these 40 words.him him him him him 27:01 >> Hello. That was excellent, Jessica, really. 27:09 >> Where are we? >> I'm right here. >> Oh, okay. 27:13 >> Sorry. Thanks. 27:15 That was excellent, and it does seem though that 27:17 in order to show incremental changes, 27:21 that you would need to subtract out 27:24 the current traffic that they have. >> That is a very good point. 27:26 >> Let's say that when number 10, and without doing anything, 27:29 they're getting a certain amount of traffic at number 10. 27:33 Then you're saying you're going to go between 5 and 8. 27:36 If they hit 8, it's only that incremental. 27:39 You'd have to subtract out, correct? 27:42 >> Yeah, so that's a good point. Where I said this was a subset of what we do, 27:44 it's not the full process, but what I can do is 27:47 we'll add that subtraction. 27:50 Oh, my heel got stuck in between it. 27:54 We'll add that subtraction into the spreadsheet, 27:56 and then we'll send it out, again, if you guys want. 28:00 You can leave your card or send us an email. 28:03 >> Right over here. Your left. There you go. 28:08 Hey. I'm curious about accuracy rates. 28:12 Have you compared this data, say, 28:16 to either the AdWords results or the Bing results 28:18 after you calculated your numbers, and then 28:21 again, to the actual traffic that was driven after you got those results? 28:23 >> The difference between us and an agency, to be very honest, 28:29 is that we're usually not involved at that point. 28:32 We get brought in to do a lot of the strategy and that sort of thing. 28:35 Since we have come up with that, I have not done that. 28:38 But what we him have done is compared this data 28:41 with the pay per click data in the preparation of the strategy. 28:45 What we'll do is remember I said sometimes the model has to be adjusted? 28:50 You use a little bit of intuition sometimes. 28:54 Like in my market, this doesn't seem to show correct. 28:57 I talked about that. That's where you'll just start adjusting. 28:59 I might have started with the 29:02 Slingshot data at the top, 29:05 but my data indicates that it's going to be different. 29:07 Then you'll just kind of bring it down, because sometimes 29:12 you look at this data and your intuition just says it's wrong. 29:14 >> Have you done anything with local prediction at all? 29:26 How local have you been able to get? Obviously, you can't track personalized results 29:29 and give an estimate on personalized, but local might be possible. 29:32 >> Like my city and state? >> Correct, yeah, and region. 29:36 >> No, we haven't done this to that granularity, 29:38 that I would think if you can narrow the search volume data 29:42 by that, you should be okay, because you're reducing 29:47 the volume significantly, so like for example, 29:50 in the Google AdWords tool, you can actually specify 29:52 state or region, so that will shrink the actual search volume number 29:56 down to represent local. 29:59 >> All right. Thank you, Jessica. 30:09 [Applause] 30:11
You need to sign up for Treehouse in order to download course files.Sign up