aka flushing money down the toilet
| 10:19 am on Feb 21, 2010 (gmt 0)|
Who was it that once said that half of his advertising budget was working well, but that unfortunately he didnít know which half? As an over-qualified, traditional-media marketing professional, now running all departments of a small business, I am embarrassed to say that I havenít a clue if any of the money I have spent on Google pay-per-click/PPC over the last 3 years has been worth it. Tell a lie, I can remember one customer who clicked on an ad and booked immediately, so I could tie things up between the two.
I am in the big-ticket service industry, where punters look around for a while before booking. Reeling them in can take ages. By then theyíve forgotten what day it is let alone what browser they used, words they typed in, ad they clicked on Ė what ad?
Needless to say, I spend as little as I can, having focussed on SERPs from quite early on. However, I do advertise, but cannot really justify doing so from a metrics perspective. I probably just feel that in some strange way that failure to do so will adversely affect my SERPs. So my campaigns are based on maximum impressions, minimum clicks - and getting quite good at it I must say!
Iím obviously reluctant to broach this subject with G itself. However, looking at it from a sales conversion standpoint, as most of my customers email me for further details, is there any way I can relate this communication to their site click in any meaningful way? Shame we cannot get IP Address from email addresses!
The sales cycle is such that it can take 4 weeks or more from first communication by email to receipt of booking form. I suppose I could then trawl Analytics retrospectively to find when an individual from a certain town clicked on my site/ad, but that sounds like loads of work.
Can anyone enlighten me as to software that might help me track PPC advertising? Any considerations on any of the above would be most welcome.
I hate to say it, but I fear Iím not only not alone in my ignorance, but very much in the majority!
| 10:51 am on Feb 21, 2010 (gmt 0)|
belenenses, sounds like you're approaching PPC in a traditional advertising way.
I too originally came from traditional corporate marketing/advertising and having spent 10 years trying to track, justify, and defend multi-million dollar print/media spends I found PPC to be a bit confusing at first but a huge opportunity after diving in.
Problem with PPC is its very benefit: conversion data. In traditional advertising one can never be sure just what drives conversion, while with PPC you can be darn sure it was X keyword on Y campaign. This means that one's work only just begins after the PPC campaign gets going--the ongoing reporting and analysis demands afterward are time intensive as well as an investment in constantly weeding out keywords/landing pages that don't work.
While I can't say that I'm as successful as I'd like to be with PPC there are some things that helps me track and judge spend success:
1. Adding campaign and keyword parameters to destination URL strings and tracking in a database.
2. Database updates upon conversion so I know what KW spent what/when.
3. Based on data modify campaigns and test new/modified landing pages.
Between Google Adwords reports/Analytics reports and your own brew of homemade tracking reports you will find enough data to make informed decisions. Key is actually getting to know your reporting options and putting them to work for your given situation.
I find canned approaches to be too generic.
| 6:58 pm on Feb 22, 2010 (gmt 0)|
Perhaps also use a web based contact form instead of a direct email link.
It can be encoded with some hidden tracking info(IPs, timestamps,refering URLs, etc.) to help tie some of the pieces together. Plus easier to record to databases for additional analysis.
| 7:28 pm on Feb 22, 2010 (gmt 0)|
If you're using PPC to drive offline conversions, tracking their efficacy can be pretty difficult but not impossible. Here are two things that might help, but both are fairly complicated.
Disclaimer: I have not implemented either of these strategies. I'm working within a more traditional e-commerce model with easily trackable conversions. However, I have studied both of these techniques and I'm sure there are many here that can speak about them more knowledgeably.
1. Session tracking - As PPC_newbie mentioned, using an online form is one way to track sessions that you can tie back to the referring source. However, it is possible to similarly track email and phone responders. If you have a "Click here to email" link on your site, you might try coding the link to fill the subject line or email address with a unique code that you can track back to that session, which you can in turn tie to a referring source. Phone calls are a little more difficult, you usually have to place something like "Refer to code ######" next to the phone number, though even then most users won't remember to use the code. Have whoever is answering the call request that code specifically, and maybe even offer the user a financial incentive or freebie for using the code.
2. Attribution - Most analytics programs use what's called a 'last-touched' attribution model. That means when a conversion occurs, the most recent referring source gets credit. So if a visitor clicks on an google ad, then clicks on a banner ad, and then bing's you and clicks on the organic listing before finally converting, Bing organic gets 100% of the credit for the conversion. Some analytics solutions allow you to divvy up this credit, either equally between all sources in a given time period, or by some more complicated attribution calculation.
Not sure if either of these will help, but there are solutions out there for business models like yours. They may be expensive and they may be complicated, but they do exist.
| 8:21 am on Feb 23, 2010 (gmt 0)|
Thanks for your comments. Much appreciated. The last two address my real issue of tracking and are food for thought.
I already have a "Click here to enquire about this product" link on each product page on my site. Thus I could instruct my site designers to code this as indicated. Iíd need to give them very precise info as they are Ďgood on design, poor on SEO.í types.
I am only doing PPC on Google, and the link takes the clicker direct to the product, so Iíd like to make sure I understand the tracking before I instruct them to do work. And I also need to get info that I can analyse and use meaningfully and in a simple format.
So if I understand correctly, I want coding that tracks the form completion & sending back to the session [days date?], and then somehow differentiates between Google SERPs and Google CPC ads?
Furthermore, as the visitor could have sent the email on a second or third visit to the product page, having clicked my CPC Google ad early in the day/week/month, the tracking needs to be able to link visitor sessions to the original CPC click. Is that more or less okay? Could a programmer in Mumbai, India understand that?
And how can this coding data be presented to me in a way thatís already analysed and thus understandable? Iím looking for something like the equivalent of management accounts as opposed to raw data.
| 5:05 am on Feb 27, 2010 (gmt 0)|
Some good info here on how to attack the problem. Personally this is the approach I take on our sites. Although we're mostly in a 2-4hour click to conversion market we still have smart shoppers that act like your customers.
My system basically works off of the sites already in place session management. When a new user arrives to the site they get a session #, basic information such as where the click came from, IP, agent, ...etc all logged. All of our PPC campaigns append a query string variable to the URL, &source=google, &source=ysm, Etc, so we can flag PPC sales in the DB. Really all we care about is where the first click came from not what happened in between so that is all we log. Once an order comes in I get a nice report showing me the order, amount, and where the sale originated from. Most of the time we can tie the sale back right there. My 2nd line of attack is done by IP. If the user leaves, loses session cookie, whatever, I can still dig up the original data by IP address. Email contact forms, like arieng said we include the session id in the contact form. If you insist on having an email link, use the mailto: format with subject line and append the SID there.
After all that we still get those mystery sales we can't match up. Many reasons still exist for this, shopping at work and purchasing at home. Simple dialup IP change day after day. Word of mouth to name a few.
| 6:25 am on Feb 27, 2010 (gmt 0)|
|After all that we still get those mystery sales we can't match up. |
Maybe from no-cookie privacy browsing. Perhaps try testing with Flash cookies with are seperate from the browsers.
| 11:56 pm on Apr 21, 2010 (gmt 0)|
"Phone calls are a little more difficult, you usually have to place something like "Refer to code ######" next to the phone number"
This is the outdated way of doing it, and as you mentioned very clumsy. Look for dynamic number insertion <snip> which dynamically insets different tracking numbers based on the ad source and even things like keywords.
[edited by: buckworks at 12:28 am (utc) on Apr 22, 2010]
[edit reason] No URLs please; see TOS [/edit]
| 4:59 am on Apr 22, 2010 (gmt 0)|
Attribution and tracking are very hot topics and there is an absolute boatload of articles and services around to help you solve the problem. You are not alone! :)
The service fuelurmind dropped (welcome to WebmasterWorld fuelurmind) is one of many businesses tackling the problem with dynamic phone numbers. ewhisper has mentioned these things to me before.
It isn't a blackhole - it just seems like that sometimes when you are starting out.
arieng's comments on last touch attribution highlight another burning debate in the PPC field.
Do some 3 or 4 word queries on Google using the terms that stick out to you in this thread and I'm sure you'll find heaps to digest. (Not permitted to drop links here)