Humbled for the Nomination of PPC Experts to follow on Twitter

January 12, 2015 Leave a comment

I’m humbled to be included in Search Engine People’s list of individuals nominated by the SEM community of PPC Experts you should follow on Twitter!

If you’re looking for a list of individuals I follow and regularly ask questions to here are some of my most often looked to individuals:

Matt Umbro – @Matt_Umbro 

Ginny Marvin – @GinnyMarvin

Kirk Williams – @PPCKirk

James Svoboda – @Realicity

Bryant Garvin – @BryantGarvin

John Lee – @John_A_Lee

David Szetela – @szetela

John Gagnon – @jmgagnon

Marty Weintraub – @martyweintraub

Matt Van Wager – @mvanwagner

Aaron Levy – @bigalittlea

Julie Bacchini – @neptunemoon

Lisa Sanner – @lisasanner

Luke Alley – @lukealley

Rae Hoffman – @sugarrae

Katy Tonkin – @katytonkin

Brad Geddes – @bgtheory

Elizabeth Marsten – @ebkendo

Sam Owen – @samowenppc

Lisa Raehsler – @lisarocksSEM

Michelle Moorehouse  – @michellemsem

Steve Hammer – @armondhammer

Christine Churchill – @chrischurch

Ruth Burr-Reedy  – @ruthburr

Duane Forrester – @DuaneForrester

John Ellis – @johnwellis

Amy Hoffman – @hoffman8

Jennifer Sleg – @jenstar

Purna Virji – @purnavirji

Bill Sebald – @billsebald

Annie Cushing – @AnnieCushing

Matt Cuts – @mattcutts

Mark Traphagen – @marktraphagen

Jennifer Sable Lopez – @jennita

Dana Di Tomaso – @daniditomaso

Eric Enge – @stonetemple

Brittan Bright – @BrittanBright

Matt Decuir – @mattbasically

Rick Galan – @RickGalan

Jonathon Colman – @jcolman

Greg Gifford – @greggifford

Jabez Lebret – @jabezlabret

Shah Menz – @michellemsem

Matt Siltala – @Matt_Siltala

Mike Arnesen – @Mike_Arnesen

Ian Lurie – @portentint

Pete Meyers @dr_pete

John Doherty – @dohertyjf

Rand Fishkin – @randfish

Danny Sullivan – @dannysullivan

Okay, so it’s not a comprehensive list, and I know that I still missed some awesome digital marketers on this list, but these are a handful of individuals I’d highly recommend you follow to get some wonderful tidbits on PPC, SEO and a few social people.

Advertisements
Categories: Uncategorized

Using PPC Audit Checklists to Prioritize Work

January 4, 2015 Leave a comment

I love what I do! I love working in Search engine marketing however it has a down side –it’s an always on space and there is ALWAYS something more we can/could do. I’m finding that some people feel like they need to work extra hours to stay ahead or just keep up to meet expectations. I’ve been digging in this last week to work on processes to share with my amazing PPC team members at Point It. Not because anyone is doing anything wrong or because I’m finding issues in our accounts, but more so as a tool to help people prioritize their day to day work schedules so they can leave work at the office at the end of the day.

I’ve made it one of my goals for this quarter to automated some of the PPC audit that we do regularly, and to establish processes to help with prioritizing work so they can challenge requests that are busy work and won’t move the needle. My end goals: reduce burn out and keep people happy.

What do I mean by establish PPC processes?

I’m working through a PPC audit checklist for weekly, monthly, and quarterly PPC Audits and tasks. Each of the audits will help uncover areas of opportunity that the account managers know where to focus to improve the performance of the PPC Campaigns that they are managing. Will it be more work? Initially yes, because whenever you start using new processes it takes a while to get used to them. However, if I can audit the PPC Audit process for Adwords via the use of Scripts it will save quite a bit of time.

Here is what I’ve pulled together so far. I’d love your thoughts and opinions:

 

Weekly PPC Checklist:

  • Campaign Budget Monitoring
  • Performance audit to prioritize which campaigns/ad groups need work
  • Bid Optimization

Monthly PPC Checklist:

  • Search Query Report Analysis  (Bi-Weekly to Monthly depending on the amount of ad spend)
  • Quality Score Audit & Pruning
  • Ad Extensions Refresh / Audit
  • Ad Copy Audit
  • Landing Page Audit
  • Remarketing List Audit

Quarterly PPC Checklist:

  • Comprehensive PPC Audit
    • Goals & Targets Analysis
    • Campaign Settings
    • Ad Extensions Audit
    • Bid Modifiers: Device/Location/Demographic/Ad Scheduling
  • Competitive Analysis & Audit
  • Specific Audits surrounding new features that were launched/released in the previous quarter
  • Develop quarterly test plans
  • Planning for upcoming launches & new campaigns
Categories: Uncategorized

Introducting Adwords Exact-ish Match

September 29, 2014 Leave a comment

Goodbye Exact Match, Hello Exact-ish Match

close-match-variants

In August the Inside Adwords blog made a quiet announcement of Close Match Variants rolling out for Exact and Phrase match types by the end of September. So what does that mean? According to Google, “Close variations include misspellings, singular and plural forms, acronyms, stemmings (such as floor and flooring), abbreviations, and accents.” Here it is, the quiet and slow good bye to Exact Match. Here are some examples of close match variants:

  1. clothes / cloths
  2. Kids / Childrens
  3. Baby / Babies
  4. Java / Coffee
  5. Cardiology / Cardio
  6. bar / barre

A few of these things aren’t like the other. Some of them make sense, some of them do not.
I’ll keep my rant to a minimum… I’m not a fan of Google continuing to take control away from paid search marketers in order to help grow their revenue reduce the amount of work to maintain large keyword lists. I understand that Google is creating a product for the lowest common denominator — small businesses who might not have resources to manage their campaigns. I get it. I really do. What I don’t understand is why Google is choosing to remove the ability to opt out. It makes me sad. Not as sad as Google forced enhanced campaigns upon all adverisers, but still sad. Google is slowly moving the control away from advertisers and professionals who know what they are doing. My solution, give the sophisticated adverisers (aka the Professionals) the ability to opt in or opt out of feature. Okay, end rant because I could write for hours. Let’s get to the meat of what this means for you.

Analyzing Close Match Variants Performance

Okay. So it’s coming. How will you know if you need to make any changes to your accounts? Simple. Analyze the performance by match type and then dig in to your top exact and phrase match terms to determine if you need to take action.
Search-Terms-Reporting


  • Login to Adwords
  • Pull a Search Query Report: In the Keywords Tab, click on details and select all search terms. (Minimum 30 days of Performance Data)
  • Put the data into a Pivot Table and segment by Match Type. Create calculations for the CPA and ROAS.
  • If the close match variants match type performance is substantially different then I’d recommend digging in by ad group and adding positive and negative keywords.



Example: For this client, based on the Phrase and Exact performance I need to go in and add some negative keywords into the account because the ROAS for Exact (Close Variants) is -50% or more of the Exact match type.
Analysis of Close Match Variance Performance

Work Arounds for Close Match Variants

Unlike Enhanced Campaigns there is a work around for the launch of close match variants that will enable you to maintain some control for your best performing exact keywords. Introducing the worlds most complex Negative keyword strategy. Just kidding, it’s not that complex, but if you would like to keep your top performing Exact Keywords as exact as possible what you can do is this:


For your top performing Exact Keyword…

  1. Create a Unique Ad Group for Each Exact Match Keyword
  2. Add all misspellings and variations as exact negative keywords
  3. In the rest of your ad groups add your top exact match keywords as “exact negative”

Here is a quick and easy visual guide for the work around for utilizing search query reports to get back to “Almost Exact” match.
Work-Around-Adwords-Exact-Match

Categories: Uncategorized

Point It Search Marketing

So I’ve got some exciting news…. I’ve taken a job with PointIt as their Director of Search Engine Marketing! I’m very excited to be joining such a great team of Digital Marketing professionals, including some very savvy women in Paid Search! You’ll be hearing more about what I’m up to and learning as the year progresses!

pointit-logo

Categories: Career

Learn about the New Google+ Interactive

November 21, 2013 Leave a comment

Learn about the New Google+ Interactive posts. They allow you to enhance your existing sharing by embedding a clickable Call to Action in your post. http://ow.ly/qWpUz

Categories: Uncategorized

Make the post to your Google Plus a bitl

November 18, 2013 Leave a comment

Make the post to your Google Plus a bitly and then link to it from your personal Facebook page. Click Promote and pay $7 to see the post. http://ow.ly/qWepd via @AimClear

Categories: Uncategorized

How to create a random A/B/n split in Excel

October 9, 2013 Leave a comment

I took over a test earlier this summer and one of the tasks I had to do was to validate the A/B/C split. I had to evaluate if my counterpart that proposed the original split had developed a random split with little to no bias between the data sets. In this post I’ll provide you with some quick and easy formulas you can use to automate the a/b/n split process as well as how to create a table that will help you know if there is bias in the split.

RANDBETWEEN Function

The RandBetween function in excel allows you to generate a random number between two variables. If you are doing a simple A/B split, you can set the fuction to =randbetween(0,1) and Excel will automatically divide half of of the rows of data to 0 and the other half to 1. If you need a three way split you can use the function =randbetween(0,2) down all rows of data you want split and you’ll see a 33% between 0, 1, and 2.

A few things to note with this function:

  1. The random split will automatically reset everytime you press enter within the workbook or make any updates. If you find a split you like, copy and paste special the variables into a new column so you don’t loose your split.
  2. The random element of the split does not guarantee that your split is unbiased. You still need to validate the split by creating a report that compares A/B/n to make sure that the variance is within your tolerance range. I’ll share a quick and simple example of how I did this so that I could hit enter until I found a split that appeared to be non-biased so I could do further statistical analysis.
  3. The randbetween function doesn’t allow you to develop a stratified split. If you need to develop a stratified split then you’ll have to embed the randbetween within an IF function and you’ll want to validate that both the split is random and valid and that the stratified variables are also evenly split between the test and control.
    What do I mean by stratified split? Check out this great article about Stratified Random Sampling by BetterEvaluation.org
    1. Evaluating the A/B/n Split for bias

      Everytime that you create a split you want to evaluate the split to see if there is any bias in the data sets. I discovered that I could use SUMIFS and COUNTIFs functions in XLS to create a table that would let me look at the output of the split to determine if the random sample appeared to be unbiased so I could then do my statistical analysis on the data set.

      Every time you enter a value into any cell on the tab that is calculating the SumIFs/CountIFs excel will randomly reassign 0,1 to the rows, giving a new potential split. Continue entering a value into a cell until the split is within your tolerance range. I try to keep my primary KPI variables of the test group within 0-5% variance to the control. A quick and easy way to visually tell if you’re split is less than 5% is to set conditional formatting for the cell background to “Green”,”Yellow”, “Red” based on the variance of test to control. Here is an example of what the A/B split table might look like and the formulas you would use to create it.


      Example of the table with the CountIF/SumIF formulas shown and conditional formatting set:
      Best-AB-Split-Formulas


      Make sure you know what your test KPIs are because it can be difficult to find a random split where all metrics are within 0-5% variance between the groups. In this example I wanted to run a test to improve my Revenue and TXN volume, so the metrics I care the most about are Revenue, Transaction, and Return on Ad Spend. My goal is to find a split where these metrics are as close as possible, and then control paste values the split before I accidentally reset it by hitting the wrong button/cell. The screenshots below are a few examples of me hitting enter multiple times to go from an okay split, to a bad split, to a better split, to the best split.

      Split 1: Revenue variance is to large. I might copy and paste values for this split into a separate column to the far right just in case after 20-30 splits I can’t get a better combination, but since Revenue is one of my evaluation KPIs for the success of my test I should keep on looking.
      Good-AB-Split

      Split 3:Bad. Notice the large variances between Test & Control metrics
      Bad-AB-Split


      Split 8: Still too much variance in Transactions
      GettingCloser-AB-Split

      Split 16: Best Split! Copy and Paste Values into a new column NOW. Notice in this split the variance in all of the core KPIs I’m going to use for evaluating my test performance appear to be within 5% of each other. Use this split for the statistical analysis.
      Best-AB-Split


      I’ve mentioned multiple times to do a separate statistical analysis post split. I’m not a stats guru so I work with a much smarter individual and team than myself, but they run multiple statistical significance calculations to determine if the split is valid. There are times when that amazing split that we saw at the end when analyzed to the mean and STD DEV are actually not valid the means are too far apart and so we have to resplit again until the means are very close together.


      I know that this won’t answer the question, “How do I determine if my split is valid?” or “How do I do the statistical calculations?” That’s not my area of expertise so I’ll try to find some links to post to help us all learn!

Categories: Analytics, Best Practices, Excel Tags: ,