How can you support this claim?  Can you discredit the possibility that the apparent improvement is just noise? How can you apply that authoritative label of “significant”?

Here I’d like to walk you through a basic test of significance that you can use to de-mystify changes in your paid search data.

If you’d like to skip the math, click here.

Let’s start with a situational example… say you’ve added Google Site Links to your brand ads and you want to show that brand click-through rate (CTR) has improved as a result.

1. First, you need to know what value brand CTR is potentially improving from.  Let’s call this value mu (pronounced myoo), and you can choose it in a variety of ways: the average or median CTR over the past month, the average or median CTR from this time of year last year, etc. It should really be whatever value you believe CTR to truly center around.
2. Next, you need data points.That is, you need several days of CTR data since the Site Links have been running. How many days is up to you. Generally, more is better, but I’ll touch on that later. The number of days you have is n. Take the average of the CTRs from those days; this is called xbar. Lastly, take the standard deviation (excel function stdev) of these CTRs and call it s.
3. Now we can compute a t-score, and with it, the probability that the change in CTR you’re seeing is or isn’t attributable to chance. Set t = |xbar – mu| / (s/squareroot(n)). Then use the function tdist in excel, and for the arguments, plug in t, n-1, and 1. The number that this function returns is the probability that the change in CTR is simply due to chance, aka noise. If this probability is very small, then we say CTR has changed significantly.

Enough Math! Is The Change In My Data Significant?

I’ve prepared an excel spreadsheet that handles the arithmetic. In this model, change the gray shaded cells to reflect your data. Enter the data that you think has fundamentally changed in column C. Only include data points since the change began. Then, in cell G2, enter the value from which you believe the data to have changed. That is, the average value of the data before the change.

The value p, produced in cell G7, is the probability that the change you’re seeing is only due to chance, and thus meaningless. Typically, a p-level must be below 5% to be considered significant. (If you want to be super, super sure, you can use 1% or 0.1% instead.) In other words, if your p-value is 5% or less, you can confidently say that the change in your data is real, definite, and due to something other than statistical noise. It’s a pretty safe bet that whatever initiative you took – whether it was switching landing pages, altering ad copy, or refining your bidding – was the catalyst for the improvement instead.

Allow me to fill in the spreadsheet with an example. For an imaginary online retailer, brand CTR hovers around 4.4%, so I fill in cell G2 with the value 4.4. The retailer enables Google Site Links, and CTRs for the 3 days afterward are 4.3, 5.2, and 5. So I enter those three data points into column C. And voila… the p-level comes back as 12.66%. This says that there is a 12.66% chance that the rise in CTR was due only to noise.

Not significant. Sorry, click-through-rates haven’t really increased, or at least, we can’t be very confident that the observed change is anything more than random noise.

But… three days is not much data. As smart analysts, we are cautious when examining trends over only a few days, and this significance test incorporates such wisdom. As the number of data points (n) you use increases, p-levels fall. For example, if all the numbers in the above example were the same except that you used 7 days instead of 3 (so n=7), the corresponding probability drops to 2.6%. In this instance, it’s very unlikely (2.6% unlikely) that the increase in CTR was due to noise, so here you can rather confidently say, “Yes, CTR has increased, and it wasn’t due to chance. It was probably due to the site links.”

• When “Statistically Significant” Isn’t Direct marketing testing is both art and science. Here are three situations where "statistically significant" results might not have business...

You may have heard hyperventilation to wit: Bing overtakes Yahoo! That turns out to be a gross overstatement, and a marvelous example of why it’s dangerous to extrapolate from small sample sets.

Here’s what we’ve seen so far in terms of PPC ads, the financial engine that drives the engines.

Remembering that Bing went live at the very end of May, beginning of June there are a couple of really interesting trends that leap off the page.

1. Totally unrelated to the Bing launch: Check out the weekend effect! Again, this is a percentage of totals, not a measure of absolute volume. It suggests that MSN gets a materially larger share of traffic during the work week than it does on the weekend. Does this mean people use MSN more frequently at work because it’s the default search engine for IE? Left to their own devises on the weekends, they choose Google/Yahoo even more often? Fascinating!
2. The Percentage of Total PPC Sales from the Big three coming through Bing increased substantially over MSN live, much more so than did costs, clicks or impressions. This could indicate better targeting logic, or that the folks giving Bing a test drive are decidedly higher quality prospects.

If we look more carefully into who’s losing the share of sales that Bing appears to have gained the answer is clearly not Yahoo. In fact Yahoo and MSN both seem to have benefited from MSN’s PR campaign. Not surprisingly the engine of choice for ‘early adopters’ — Google — is the one seeing its user base sniff out the competition.

With respect to Bing, the initial uptick in the quality of traffic is quite encouraging. Higher Sales per Click will lead to higher bids and more monetary share for MSN. The fact that despite the media blitz surrounding the launch, the impression share and click traffic share hardly moved is somewhat more concerning. If consumers are truly happier buying through Bing it could give MSN some momentum, but they can’t just rely on the early adopters and search geeks. They need to pull in and keep the average Joe as well.

As Lance Loveday points out, Bing has a steep hill to climb and it’s not clear that the media approach they’ve taken is going to do the job.

We’ll keep an eye on the trends. MSN has finally given us something to write about!

A follow up post, eliminating trademark search and controlling for a couple other factors paints a very different picture.

• Bing: A Closer Look at the Numbers A more careful study yields a different perspective....
• Bing: The Art of Buying Share It's hard to buy loyalty....
• Bing Surges: Google buys TV ads Bing came out of the gates hard in 2010!...

Does your PPC program drive more sales than it’s credited for? The answer is certainly “yes”, but the sources and volume of under-reporting might surprise you.

If you drive your program through your web analytics software you may be missing 10 - 30% of the sales! This is not the fault of the software as much as it is the convenience of using javascript tags for tracking. In past years, part of the problem came from browsers not happy about running 3rd party javascripts, but those problems have been fixed.

The problem now is not with javascript flakiness as much as when the script is run. Almost everyone wisely puts the javascript tracking in the footer of their web pages. The reason? You don’t want the customer to have to wait for the javascript to run before loading the rest of the page. If there is any problem, you want the page to load and the footer to hang so that the user can shop unimpeded.

Herein lies the rub. Because it’s in the footer the javascript can only cookie the browser after the whole page loads. For heavy, slow, image-laden pages, customers often move on to the next page before the footer loads. If the user sits on that subsequent page long enough to fire the javascript the problem will be that you’ve lost the url parameters that allow your tracking system to know the source of the traffic. That user is now flagged as “untracked” even though they came through a paid advertisement.

We know this happens and understand the scale of the problem because it shows up whenever we do data audits with clients, but also because we sometimes employ both our standard tracking and a javascript tag when we’re studying marketing channel allocation for our PPC clients. The problem isn’t that orders we see are tracked to other programs; it’s that the sales we know came through a PPC ad aren’t tracked to any marketing program.

Knowing this, you might say: “Well, if I have a sense that this happens 20% of the time, can’t I just adjust my advertising efficiency thresholds by 20% to compensate?” Yes, you can, but the problem is that some destination pages are more susceptible to this problem than others, either because they load more slowly, or because users are more likely to navigate off of them quickly. This will disproportionately penalize some types of keywords over others resulting in lost opportunity as those ads are mistakenly bid down the page.

A better way to track high-dollar marketing programs is through use of a fast redirect. The redirect is fast if, and only if the redirect server doesn’t have to do a database look-up. If the server has to look up the destination url the redirect will be slow and the server will bog down during traffic bursts. We pass the final destination url to our redirector as an encoded parameter so the redirect takes less than 0.1 seconds and the volume of redirects is almost irrelevant.

Using a redirector provides much more robust tracking, but can/should be cause for concern as well. With all of that valuable traffic passing through a third party box it’s valid, indeed essential to ask: “what happens if the box goes down?” We stressed out about this, too. Our approach was to build in multiple redundancy by having multiple redirectors. To keep these independent, these servers are located across the country, and use different internet backbones. All the servers share the work, and are self-checking and self-correcting. If a data center becomes unavailable — say, due to a server failure (almost never), or due to DC connectivity problems (very rare), or due to routing hiccups somewhere on the web (not rare, but brief) — we use smart DNS to reroute traffic to the healthy machines within a minute. No system is 100% perfect, but this redundancy and automatic checking provides extremely high uptime.

Handling order allocation issues can be done on the fly, with the confirmation page tag sending marketing allocation along with the order details, or through a back-feed of order ids and marketing channel credited. Any competent search agency can then base its bidding and reporting on only those sales your system has flagged as paid search or unknown.

We see all the costs associated with PPC advertising, but we don’t see all the sales generated. While this post may cause some eyes to glaze over, understand that this is not minutia by any stretch of the imagination.

Better tracking technology plugs a big hole, but others remain. Users drop cookies, use multiple browsers, and sometimes search on one machine but place the order on another after shopping around. We can track spillover to the call center, but measuring foot traffic driven to the stores remains elusive. However, those who throw up their hands and conclude that direct marketing metrics shouldn’t be applied to search simply aren’t trying hard enough.

This is isn’t a big deal for online retail yet, but it will be, in the next 6 to 18 months.

Today, HTML markup is about how data should be presented. By tagging elements on a HTML page with standard class names, microformats are about what data mean. Microformats will form the foundation of “Web 3.0″, aka the Semantic Web.

Google is starting with microformats for people (hCard) and reviews (hReview).

It seems clear Google support for the SKU microformat (hProduct) is coming soon, as there’s already documentation for it on Google WebMaster Central.

What will this mean to online retailers long-term?

• The first retailers to add hProduct tags to their product pages will enjoy an early sales advantage, as Google will present their products first because of the extra data.
• The advantage will be short-lived, as large retailers and standard e-commerce platforms will quickly jump on the band wagon.
• Retailers will chafe at the simplicity of the format. hProduct doesn’t encompass shipping, tax, bundled pricing, or even UPC. Argh. When Google miscategorizes or misprices your product atop their SERP, get ready for additional customer service calls.
• hProduct markup is essentially a product data feed, albeit with limited fields. Google’s current product search hasn’t gained much traction, but widespread hProduct data will help Google disintermediate the shopping comparison engines.
• Long term, hProduct markup will increase consumer perception of Google-as-store, eroding weaker retail brands (see “search engines atomize retailer brands”, halfway down this post.) Building a strong well-defined brand is key.
• When Google presents more SKU-centric multi-merchant data on the top left of the page, paid search on the top right becomes more crucial to get your link in front of shoppers. The paid search core competencies –optimal bidding, extensive keywords, and solid “why-shop” copy — become even more important.

What should online retailers be doing about this today?

I’d recommend keeping a close eye for mentions of hProduct on the WebMaster Central Blog.

I’d also recommend planning to add hProduct tags to your SKU pages, probably by late ‘09 or early ‘10. (Here’s the link the Google hProduct documentation again.)

For better or worse, this is where online retail is heading, and your marketing and merchandising teams will benefit from being there at the beginning.

• Online Retail Stats: CyberMonday 2008 vs. 2007 One day isn't an entire retail season, but I thought it interesting to compare CyberMonday 2008 to 2007....
• Eight Important Online Retail (and Marketing) Policy Issues Scott Silverman of Shop.org talks about internet privacy, interstate collection of sales tax and other issues facing online retailers....
• IsabellaOliver.com Overcomes 3 Key Challenges of Online Retail Videos Because they overcome the typical (and non-trivial) challenges with such panache, it's worth your time to check out to the...

This past weekend a couple of different incidents made me think about the source of customer loyalty.

On Saturday afternoon there was no hot water at my house. In my vast experience with home improvement projects I’ve learned that three things generally happen when I try to fix something: 1) I curse a lot; 2) there is bloodshed; and 3) I end up calling someone who knows how to fix things.

I checked to make sure the circuit breakers weren’t thrown, that there was voltage at the unit, I pushed a “reset” button. Nothing worked. Sunday afternoon I gave up and called Curtis Heating & Cooling. I’ve used Curtis’ for years for HVAC stuff and was delighted to see that they worked on water heaters, too.

They suggested over the phone a couple of things to try but I’d tried those. I suggested that they’d probably need to replace the unit and he said: “well, if we do, I wouldn’t be able to fix it today anyway, and I’d have to charge you more for the visit since it’s Sunday. Can it wait ’til tomorrow?”

Curtis Morris, himself, came out first thing Monday morning and spent an hour or so diagnosing the problem. The problem was electrical after all, the circuit breaker switch was bad and wasn’t providing enough current for the water heater. He suggested we call an electrician he knew as he wasn’t certified to do that kind of work. Curtis wouldn’t accept payment for the hour he spent at our house.

This is the third or fourth opportunity Curtis Heating & Cooling has had to charge me thousands of dollars to fix a problem I thought I had, and instead provided a simple solution that was more than adequate. It’s why I will always call them, and why I tell everyone I know to call them. {A colleague of mine was talked into replacing his heat pump to the tune of $3K by another company. I suggested he call Curtis just to be sure. Curtis fixed his system for $73.}

By looking out for his customer’s interests first, his company suffers in the short term but benefits ultimately by creating customers for life.

Also this past weekend I ran across a blog post by one of RKG’s former IT Developers. He left to help build a company that provides open source software development. His article touts the merits of charging fair hourly rates instead of billing clients for on-site visits and work they don’t need, suggesting that at the end of the day the strongest business model is to provide valuable service for a fair price.

This is a model we’ve tried to follow at RKG. Where some firms seem to see themselves as commissioned brokers for the engines, we take the view that focusing strictly on what serves our clients’ best interests — regardless of near term impacts on our bottom line — will lead to long-term growth and prosperity.

As the economy recovers, and corporate survival is no longer the only issue, I think we all do well to reflect on the lessons taught by Curtis Morris and consider what we can do to engender that kind of loyalty.

• Fred Reichheld Discusses the Importance of Customer Loyalty to Business Growth “The only way to grow your business long-term is through this process of turning your customers into your sales force.”...
• PPC Engine Loyalty Study A new study on search engine loyalty from the RKG skunk-works...
• Customer Service Lessons From The Heart Interesting post from Mike Wagner on the perspective into customer service gained from days in a cardiac rehab unit....

As part of this approach we’ve launched “truth campaigns” to debunk various myths that are propagated in the industry by those who follow a different business strategy.

The PPC buying cycle is one of those notions we’ve tried to put in proper perspective. The buying cycle argument suggests that people first search for general product categories, then get more specific later as they get closer to placing an order. Its advocates then argue that the more specific keywords steal credit from the more general keywords that got the ball rolling, and that it is therefore a good idea to spend well beyond your efficiency targets on general keywords, because while it doesn’t appear that they’re driving enough sales to justify their costs, the sales those keywords drive are simply being credited elsewhere.

There are two fundamental problems with this line of reasoning:

1. This pattern is far less common than proponents suggest. Some less-than-scrupulous agencies have convinced their clients that changing the efficiency thresholds for these “early stage” keywords by factor of 200% or more makes good sense. We’ve seen instances where retailers with 30% cost to sales targets were convinced that a 300% cost to sales ratio (a 1000% increase!) on these keywords was acceptable.

   The reality is that even giving full credit to the first ad touched rather than last would rarely move the needle more than 10%. The highest we’ve ever seen is ~40%, which would mean moving the efficiency target from 30% to 42% on that KW.

Here’s keyword data we pulled for one of our clients (we’re actually bidding to margin for this client hence the varied cost to sales values):



2. Second, it’s not clear how much credit early stage touches actually deserve. There are several grounds for questioning the value of that first touch.
   • In some cases, the initial search is in a completely different product category from the subsequent search (first search: “Cuisinart”, second search “ipod nano 8 GB”). We investigated a sampling of these mismatches and in each case the purchase was related to the second search, not the first.
   • If, after the initial visit, the customer still hadn’t decided where to shop, and was willing to see a wide range of retail offerings for the specific type of product he or she wants to buy, exactly what was the value of the first visit? Why do we care whether the initial process of narrowing choices happens on your site (and on your dime) or your competitor’s?

Perhaps there is branding value to the initial, unproductive visit. Perhaps, people who shopped around on your site are more likely to click on your link when the do the follow up, more targeted search, than are folks who bounced around someone else’s site.

If we then are willing to give some credit (10%, 20%?) to the first PPC click how does that affect the math?

Now we see that if giving all the credit to first rather than last only moves the needle 5 or 10% most of the time, giving 20% credit to the first click instead of 100% only raises your efficiency threshold 1% or 2%. In our example before the retailer could afford 30.3 - 30.6% on these early cycle terms rather than 30%. In our most extreme case giving 20% credit on the keyword that saw a 40% lift when all the credit went to the first touch, our efficiency target go from 30% to 32.4%.

These micro tweaks are unlikely to have much impact on bids and positioning on the page, hence may have zero impact on traffic.

Worth investigating, sure, but if someone tells you that a huge difference in efficiency targets is warranted and profitable based on the buying cycle they’re either ignorant or lying.

I’ll take up the topic of branding value in a subsequent post. I’m wrestling to understand how branding is justified when tracking proves it to be unprofitable…

Love to hear other thoughts, opinions and data on the subject!

• The PPC Buying Cycle: Buyer Beware The Buying Cycle is used by some as an excuse for overspending on general keywords. It's a lovely theory, but...
• Thoughts On Multi-Channel Strategy Scratching your head over the interaction between your online and offline marketing efforts? ...
• Top 10 Paid Search Posts Top 10 Paid Search Posts Visit RKG at Booth #352 at the Internet Retailer Conference At Our Booth: We’ll be...

A couple of notes on the data:

• The benchmark we’re using is advertising expense not impressions or traffic volume.
• We looked at Paid Search only, not content or banner deals
• We studied the advertising network, not strictly the engine’s domain.
• Our systems are engine agnostic: our clients do not set engine budgets, indeed most don’t have budgets at all. We allow money to flow based strictly on where traffic can be purchased within efficiency tolerances.

Given that understanding, here’s what we saw over the past 15 months:

Two observations come to mind:

1. It appears that since last August, Yahoo! has regained some ground.
2. Google seemed to get a larger holiday kick than the others.
3. Even with MSN Cashback and MSN’s willingness to buy Market Share from Google — I’m sure others saw Google ads for Live Search on high traffic KW, yes? — MSN gained no share whatsoever.

With respect to 1, I’m not sure if the fluctuations between Google and Yahoo relate to Yahoo showing Google ads for a period last spring and summer, or network partners coming and going. Could Google’s holiday bump be a function of advertiser’s focusing more on Google bidding than others at the holiday driving up CPCs for those of us who apply seasonal adjustments automatically across all engines?

Do others see similar trends? Do others have better explanations than I do for them?

• June 2007 Paid Search Market Share: Google, Yahoo, Microsoft Google continues to comprise three-fourths of our agency's total monthly PPC spend. In June, Yahoo held steady, while Microsoft picked...
• Google, Yahoo “Panama”, and Microsoft: Paid Search Market Share May 2007 We review May 2007 ad spend data for our agency. Google continues to dominate, with three fourths share. Google lost...
• June 2008 Google, Yahoo, Microsoft Paid Search Market Share Looking at our agency's client base in aggregate, last month Google received 79% of our clients' ad dollars. Yahoo received...

Today, I get to participate.

I noticed a new feature on my Google SERP for all of my searches this morning. An extra “web” plus box in the upper left hand corner, under the Google logo.

Expanding the box gives the user an opportunity to further filter their results by recency (day, week, month, year) or by type (videos vs. forums vs. reviews). It also provides the ability to add images from the websites that appear in the results or to add widgets that allow you to see the SERP in a different view.

Of the views, “Wonder Wheel” seems to be the most novel and the least helpful. Changing the view to “Wonder wheel” moves the traditional organic listings to the right and create a spiderweb-like display of related search terms. Clicking on a more specific term pulls up another web with more deeply related terms.

The plus box appears regardless of whether I’m logged in or not and whether I allow the SERP to be tailored to my locality or not. Others around there aren’t seeing this, although the folks over at Blogoscoped picked up on these changes and even provided a way for you to check them out.

Officially, Google did announce two updates yesterday regarding longer snippets and their implementation of Orion. (Greg Sterling also posted a description and examples of those changes over at SearchEngineLand.)

Has anyone else been “selected” to test these features? Any thoughts?

• Test Driven Development At Google Google advocates for test driven development, no matter where an engineer sits....
• Google PlusBox Performance Results Reaction to 6-months of Google AdWords' PlusBox performance data....
• Google AdWords Plusbox Revisited New Google AdWords Plusbox Functionality, SERP Layout & Notable Data Updates...

There are many different types of Paid Search agencies. Some rely on the engine’s UI and spreadsheets to do their work, the more sophisticated have either built or rented a tool set to use for bidding, ad management, etc. Some in this last category expose their UI to their clients, while others, like RKG, use their proprietary tools on behalf of their clients.

The tool providers call around and offer tours of their products and I usually take them up on it…once. It’s always interesting to see where others are in the evolution of the industry. Most of these tools do the basics very well and would be a huge help to an agency just wading into the space.

That said, on the all important analytic front the best of these systems is about where RKG was in 2005 — not bad, but nowhere near where we are today. Sometimes they ask: “what would RKG need to see in this tool that would make you seriously consider switching?” “A reflection of my corpse.” is my honest, but not so helpful response.

Without going into the underlying statistics, the data architecture, methodology and other important differences the biggest weakness is perhaps in the limitations of the UI. I’m not talking about specific features that are missing, I’m talking the limitations inherent in any UI.

I’ll show you a screen shot of the most powerful piece of RKG’s UI:

Yep, it’s just a blank query window. In that nothingness, our analysts have the flexibility to find and do anything and everything. The power to custom-craft any data, any analysis, any bidding enhancement to fit just the ads and data you intend cannot be hard-coded into a UI.

The drawback to infinite flexibility and that power that comes with it is that we can’t expose our UI to our clients, and we have to hire remarkably sharp people who can learn to be power SQL users and train them. We’ve had to grow slowly relative to others — still 315 on the Inc 500 list, but — and only recently moved from having one full-time sales person to having 2.5 FTEs on Marketing/Sales/PR.

We think the benefits far outweigh these inconveniences. As I’ve argued before, you can’t lead in this space if you’re using the same core technology as others; the ability to add features and tools as needed is irreplaceable, and the power and flexibility gained by having thoroughly trained sharp analysts with the power to code anything they need into the system cannot be supplanted.

It is easier to follow a different path, but we think we’ve found the right one for our firm.

• Naming is Power 2: Being first vs. being big. Early movers get to shape conventions. If you show up late, it helps to be large....
• Google, General Electric, US Dept Of Energy Announce Open Platform for Power Generation A broad alliance of leading technology and power generation companies today joined forces to announce the development of "Zappy", the...
• “Power 150″ Top Marketing Blogs Honored to make to the Power 150 List of top marketing blogs....

We ran an analysis to show the year over year performance trends over the last 7 months or so. We took data from our 40 largest retail clients, and tracked the median of the year over year performance differentials in several different areas. For comparative purposes, we simply divided this years’ numbers by last year’s. 100% represents the previous year’s performance level that week. As the graph shows, the numbers are discouraging.

While a number of our clients asked us to sacrifice efficiency for top line sales in the early part of the downturn, most have chosen to pull back at this point and protect the bottom line.

We hit a wall around Week 37 of 2008 at about a 20% Year over Year drop-off in sales and costs, and those numbers have gotten slightly worse since Christmas.

Let’s peel the onion to figure out exactly how the economic downturn has impacted paid search.

Pretty clearly there are three possible mechanisms that contribute to a decline:

1. Decline in Sales per Click: if a smaller fraction of visitors through PPC ads make a purchase or the revenue per sale drops, or both then a retailer will generate fewer sales.
2. Competitive Pressures: if declining sales per click force an advertiser to lower its bids more than its competitors, the advertiser will drop down the page and get fewer clicks per search.
3. Fewer searches: tightening wallets may affect user behavior in another way. In addition to potentially lower Conversion Rates and AOV, one may well see fewer searches in general.

Take a look at the following graph showing YOY Conversion Rates and Average Order Values on competitive search. For consistency we’re just comparing Adwords to Adwords.

There’s some fascinating stuff here. First, note that late in the Holiday season, AOVs were pretty close to ‘07 numbers for most of our clients, but the conversion rate suffered. Since the start of the new year the problem has reversed: conversion rates are about where they were in early 2008, but the AOVs are off 10% or so.

Revenue per visitor is part of the problem, but by itself it’s not the whole picture.

Could it be the competitive landscape? Are our clients having to pull back more than their competitors hence getting a smaller share of what’s available?

This suggests that our clients were about half a position lower than the previous year for much of the holiday season, but are actually higher on the page than they were last year at the same time. Averages lie, and medians of averages are a pretty squishy measure, but this suggests to me that our clients (and our analysts) have reacted appropriately to the landscape and efficiency needs, and that they’re not totally out in left field compared to their competitor’s reactions.

Competitive pressures may have impacted Q4 slightly, but don’t explain the YoY drop in Q1 at all for our client set.

This leads us to the last explanation: traffic volume. Are there just fewer people out there searching for our clients’ goods and services? Unfortunately this is among the hardest factors to gauge. Because of match-type and network partnership shenanigans, impressions aren’t an accurate measure of anything. Click volumes are impacted by position on the page, so that’s not the right measure either.

Instead, we picked “brand” traffic as a proxy. Figuring that our ads are in position 1 on our client’s trademarks always, the traffic and sales volumes coming off of brand search may indicate general levels of demand as well as anything. This is far from perfect. Probably very far. Brand search is largely a function of other offline marketing efforts as well as loyal customers, so as our clients pull back on those efforts Brand sales may fluctuate. Also, customers may be more inclined to latch onto those organic links to affiliates promising discounts on retailer’s brand names.

With those caveats understood:

This suggests to us that the main cause of decline is simply fewer people shopping. The people who are shopping are spending less, but the biggest dent has come from traffic.

We remain optimistic that these numbers will turn this summer and we’ll start seeing growth again in Q3, but it could be a rough ride between now and then.

We wish everyone the best of luck in these trying times.

• Barron’s: Y! possibly more valuable as G affiliate? Ouch. Eric Savitz over at Baron's wrote a short provocative article yesterday. To whit: Y! could increase shareholder value by reverting...