Ask Your SEM Agency These 4 Questions


How confident are you in your SEM agency?

There are some sharp folks in the SEM space, for sure.  Then, why is it that companies seem to change their SEM agency every two or three years?  While I'm working on AdWords reviews, I can see the lineage of past agencies piled up in the change logs like layers of different rock in a canyon.  I've even had the chance to see before-and-after views in some cases.   Some results were pretty impressive--I built up some solid respect for some of my paid search peers.

So why doesn't all of the PPC business pool up under one brilliant SEM agency over time?

The challenge is that each of these agencies have certain strengths.  They come in, fix the things they're especially good at, and then plateau.  Agencies and consultants aren't only replaced for outright poor performance--they're more often replaced due to limited continued growth.  It's very difficult to pull the same rabbit out of the hat twice.  "They say they've done pretty much all they can do at this point" is what you'll hear from the retailer at this stage.  And then they're willing to consider alternatives that might get them to the next level, hopefully while preserving the gains they had under the previous care.

When I was with PoolSupplyWorld, and we started seriously focusing on AdWords, I added 30% to the business in the first year.  Growth was solid the second year for that channel, but it wasn't 30% again.  Why not?  Each major advancement isn't necessarily a change in trajectory--it's a change in elevation.


Ask Your SEM Agency These 4 Questions - Actual Result Chart


In order to continue to make incremental gains, you have to do new things.  Each company only has a finite set of tricks up their sleeve, so the only way to identify potential longevity in a partner is to see how many tricks they have.

There are some common corners that AdWords Accounts can paint themselves into, though, and there are some questions you can raise that might shake it loose.

1. "How profitable are our Non-Branded Campaigns?"

I covered this topic in more depth in this article, but the premise is this: if your Branded and Non-Branded Campaigns are lumped together, then you could be actively losing money in your Non-Branded Campaigns, and it would be covered up by the apparent strength of the Branded Campaigns.

If you aren't providing profitability information to your Agency, you could reframe this as a question of ROI (or equivalent metrics, like COS or ROAS). "What's the ROI of my Non-Branded campaigns relative to my overall ROI?"

If your Agency is taking advantage of the lack of transparency afforded by Account-level reporting, they'll be squirming at this point.  But if they come back, and tell you that your Non-Branded Campaigns are performing at a ROAS of 8x, and you know you need 10x to break even, then you've definitely uncovered an opportunity for better performance.

If you'd like to check for yourself (and I recommend you do), simply go into your Account, and click the "Columns" drop down, just above the graph.  Then, add "Conv. value / cost" as a column.  This is, as far as your AdWords conversion tracker can tell, the ROI of your Campaigns.


Ask Your SEM Agency These 4 Questions - Adwords Modify Columns


You can then see your ROI per Campaign.  Anything stick out?  Anything wayunder what you were hoping to see?  You can aggregate these values by using a Filter.  For example, if your Brand campaigns include the word "Brand" in the Campaign Name, then you could do this:


Ask Your SEM Agency These 4 Questions - Adwords Filter


At the bottom of the table of data, there will now be a yellow row, titled "Total - all filtered campaigns".  Under the "Conv. value / cost" column you added earlier, you'll find your Non-Brand ROI.  If it's a surprising value, then you may have uncovered a great opportunity to improve the return on your account.  Congratulations!

2. "How are we applying Bid Modifiers?"

Bid Modifiers are coefficients that AdWords allows you to use to adjust bids across some other variable.  Let's take a look at the three most common--Location Adjustments, Mobile Adjustments, and Date-parting.  All three frequently reveal the same mistake.

The "If It's There, We Should Use It" Mentality.

The common mistake here is simply overzealous over-use of the tools.  AdWords makes it very easy to see the relative performance of different geographies (through the Dimensions Tab), and across device types (through the Device Settings of the Campaign), but it provides no tools for determining statistical significance.  As a result, many account managers will set adjustments based on an arbitrary time window, regardless of sample sizes.

In theory, this should even out over time, as decreasing bids should act like a "price skimming" model, where the clicks you're still attracting are the highest return.  On the other side of the model, if you bid too high, then the collected value per visitor will drop, and bids should then settle back down.

Unless you're chasing shadows.  This methodology only works under two conditions.  First, you have enough data to determine that you have a statistically significant difference before you decide to apply the modifier.  Your Agency should be able to describe to you how they determine whether or not a modifier is merited.  Second, this methodology only works if regularly re-applied.  How long has it been since someone looked at your modifiers?

All three types of modifiers are subject to the same requirements.

A common refrain is "well, your ads are doing better between 3pm and 5pm, so we should increase bids then."  Really?

As it turns out, a difference in volume is not the same as a difference in quality, and only the latter justifies a change in bidding strategy (outside of some fringe cases involving tight budgets).   The fact that you get more clicks or conversions during a timeframe, or in a particular geography, or on a particular device type, does not mean that you should increase bids for that segment.

If you're bidding with the intent of producing revenue at a particular ROI, then only changes in AOV and Conversion Rate should affect your decision to apply a bid modifier.   This can also be detected as a difference in Value Per Click (VPC), which has certain advantages, as a KPI.

If you're happy to spend 10% of revenue on ads, then why would the time of day matter, unless it affected that outcome?

3. "Beyond just optimization, what's your plan to build new business?"

As a retailer, this was one of my favorite questions.  As a service provider, now on the other side of the fence... it's my favorite question, period.  A client that asks me this is thinking about their account in a very practical way, and is putting the ball in my court--exactly where I like it.

This Is Where a Stale SEM Agency Partner Will Shrug

It was this realization that finally explained to me why I was seeing such agency turnover within paid search.  Too few providers were challenging themselves to continue to improve.  They'd cook up their strategic advantage, and then their sales teams would go out and find as many people as possible to buy it.

But once you buy it, and have it, what's next?

This should be the most fun part of the conversation between a provider and a client, and Agencies that shy away from it, waving their hands impotently, are a red flag.

4. "Do you mind if I have XYZ perform an Account Review?"

Want to test the confidence of your SEM agency partner?  Ask them if they'd be willing to look at someone else's account AdWords review.

An SEM agency partner who is open to new ideas, and is willing to help your account find incremental gains will jump at the opportunity.  After all, from their perspective, some other company is going to pay some staff member to go through the account with a fine-toothed comb, and there's nothing stopping them from taking advantage of that effort on your behalf.