PPC Account Setup

Posted on February 1, 2016

Kirk Williams recently published a post on PPC Hero about PPC account setup for ecommerce clients. It is a good post and you should read it; the main takeaway for me is a reminder of how many things it is possible to do in an ecommerce account. It also made me think a bit because my approach to a new account build is different; this post explores which one of Kirk or myself is wrong. Spoiler this is an “it depends” situation.

Firstly, what is my approach?

As with most things PPC, I don’t have a process but instead a set of principles which can be applied to generate something that looks a bit like a process from the outside. This is, in some ways, a major weakness of my practice and would be something that would need urgently fixing if I were to try to scale my one-man-band consultancy into an agency.

1 Account Build Principles

1.1 An account is never finished

There is no point at which a PPC account is “done”. There is always more keywords that could be added, more ad groups that could be segmented, more ad extensions that need to be tested.

So what does finishing an account build mean? It doesn’t mean the account is built. Or at least it doesn’t mean it in the sense that a house or a tower block are “built”.

I think the way most people use it, an account build being finished means that the account is now in a state where it is expected to achieve targets whilst being managed in “maintenance mode”. Maintenance mode doesn’t mean that the account is being neglected, just that the work required is much less intense.

1.2 The main constraint is my time

There is only one of me and I have other clients. So I can’t just drop everything for a week and work purely on one account. And I have too many fun things to do that don’t involve AdWords to willingly do all my normal work and then overtime on an account setup.

Therefore, for me, the hours spent on an account setup are spread out - possibly over a number of weeks.

This is a necessary constraint of my business model and personal priorities. YMMV.

1.3 At the start, no one knows what is best

I’m not a fan of the “you must test everything” and always assuming a position of complete ignorance to start with; we all know that brand keywords are going to have a better ROI than shitty placements on the GDN. So please don’t take my “no one knows what is best” phrase too far.

But there are limits to how much I know, especially when starting out with a fresh account is a vertical I haven’t worked in before. What will get me the best ROI on time? Should I be adding more keywords? Or should I be working on better segmentation in the shopping campaigns? Both of these tasks are likely to make the client more money - but which one is the best?

Given the time limitations (principle two) and that the account build is never finished (principle one) this means that the main challenge is figuring out what to do, what not to do (or what to postpone) and the order in which things should be done.

2 How does this turn into a process?

My duty to my clients is to get them the best return on the money they invest in me. N.B. this can be quite different to getting the best ROI for the account.

The first point to note is that there are some circumstances is which the “return on time” of an action depends very strongly on when the action is done. For example, if it is Valentine’s day next week then the return on getting the Valentine’s stuff ready is going to be pretty high if you do it right now and pretty low if you do it next month.

So first, identify time critical events and make sure that those things are done on time.

Apart from events driven campaigns like this, there is not a close link between when an action is done and the return on it. It is better to get things done sooner - a campaign making £1000/week will make £2000 if you get it done now rather than in two weeks time - but limits on time mean that not everything can be done instantly.

The point is that, once the event driven campaigns are sorted, the order in which actions are taken doesn’t really change whether I have 40 hours a week on the account or four hours. In the latter case, things just take ten times as long.

Now we are in the position of trying to forecast the return for campaigns we haven’t made yet. Some readers will say this is impossible, but most are doing this all the time in an informal way whenever decisions are made about what to do next.

I can have a guess about what kind of things will see a good return. Taking into account the amount of work needed, the priority list will look something like this:

  1. Brand keywords - incremental return isn’t huge, but easy and quick to get started with
  2. Top category keywords - the direct return on these keywords is likely to be pretty tough to start with. What we are really looking for here is information which will be pretty important for the next steps.
  3. Shopping - as long as they have a product feed, this is easy to get started with

This is quite a short list. Compared to Kirk’s action points, the scope of work and the completeness of the account is extremely limited.

But we are just getting started!

It should be possible to get basic versions of the above things live within a few days (certainly within a week) and then a lot of the fog around prioritising actions goes away; there is data!

Immediately (or after a few days), you can start making decisions on which categories to expand first, whether it is more important to do more keywords or more shopping, or if it is vitally important to work on ad extensions. I’ve even had a client where it was apparent at this early stage that search volume was not high enough in their niche to cover the cost of managing the campaign; in this case we were able to redirect my efforts into something else which saved me the embarrassment of billing for work that wasn’t going to make them much money.

There are clients that I’ve worked on for months where we are doing no DSA. This isn’t because I think DSA is bad or that it won’t work for them - I think that it won’t work for them as well as other things that are higher up the to do list. And, luckily for me, they trust me to make these kind of decisions. (There is a case here that these guys may be better served by a larger agency who can give them more time, but that is a question for another day).

3 So is Kirk wrong?

Firstly, I would never say Kirk was wrong in this area - I don’t know enough about his clients or business to say what makes sense for them and him.

I know that Kirk charges a setup fee for an account build which sounds like it is a fixed fee charged for a particular project. For fixed fee projects, it is extremely important to have a scope of work agreed in advance.

The other thing I know about Kirk’s business is that for ongoing management he charges a percentage of media spend. This means that the monthly fee could be too low to cover the build out of new features without further negotiation - it certainly means that the risk of building out new stuff is on Kirk as if it doesn’t work out, the client won’t put media budget through it and his time will have been wasted. Seeing as further negotiation on fees to cover new build activity is mega awkward there is an incentive (on both client and agency side) to move extra work into a separate line item on the invoice. In this case it absolutely makes sense to take a “I’d like to build everything straight away” approach because otherwise there might not be the opportunity to do so later - this is bad for the client and bad for the agency too.

My model is different - for AdWords management I charge a day rate with the number of days being agreed in advance. This means there isn’t really a difference in billing between speculative time spent exploring new build opportunities and time spent on more mundane management. Under this model, there is only one advantage to having a fixed scope build project over my preferred “just take the best action at any given time” approach; a fixed scope build project guarantees my availability and delivery of the project in an agreed time frame.

For the clients I work with, this doesn’t seem to be an issue, but I can believe that there are marketing departments out there for whom this is a key requirement. This could be another example of me not really having much of a business plan.

When I first read the post on PPCHero I was surprised that Kirk’s methodology was so different to my own. But after further thought it comes down to the old question of how is it best to bill clients.

I have a post brewing on this question, but it will be a while before it is ready for publication.

4 Reaction and Addendum

There were a few reactions to this piece that require a response or some clarification:

@RichardFergie @PPCKirk What you’re trying to say. Basically what I got was, most of the points in Kirk’s article takes too much time?

— Sean Burrows (@BurrowsPPC) February 1, 2016

@RichardFergie @BurrowsPPC at some point, have to build out Shopping/Search & only way to is to just do it. Might as well in beginning IMO

— Kirk Williams (@PPCKirk) February 1, 2016

Dealing with the first point first: my point isn’t that specific things take too much time; my point is that there will never be enough time to do everything that it is possible to do.

This means that there is a task prioritisation problem - you must decide which out the the millions of actions you could take to actually do. This problem crops up all the time in PPC.

My view is that it is easier to make good prioritisation decisions with a bit of data. So it is better to get some simple stuff live as soon as possible - the data from this can then be used to inform what you do next. This might mean that you don’t do some things at all - if there is more money to be made adding keywords then why waste time with retargeting until the keyword adding hits diminishing returns?

I think my argument is exactly the same to Kirk’s point. “At some point, have to build out Shopping/Search” is true, but it is also true that both these things are never fully finished. There will always be more things to do.

So again, it is a question of prioritisation.

  • Which categories is it worth spending more time on?
  • Is it better to look at more keywords or more shopping segmentation?
  • etc.
  • etc.
  • etc.

It is important to note here that you can’t avoid prioritisation decisions here. By committing to a scope of work defined in advance, all these decisions have been made in advance.

Making these decisions as early as possible is the best option - provided they are the right decisions :-) And an experienced PPCer, operating in a vertical they know well will probably get this sort of thing right most of the time. So I wouldn’t expect a huge improvement in performance by doing things my way.

Finally, we can look at the timeline for an imaginary build project which will, hopefully, make my point clearly:

Time Flexible Scope Fixed Scope
Start Agree number of days. Figure out when I can fit in those days. Come up with timescale Agree scope. Figure out how long it will take me to implement scope. Agree finish date
1 Basic brand campaign to go live, start work on top product categories Basic brand campaign, start work on top product categories
2 Product categories go live. Quick shopping campaign setup (submit feed, create “all products” campaign. Continue product category work
3 Shopping campaign goes live. Review product category data Finish product category work
4 Expand top product categories. Neglect product categories with low potential Start work on shopping campaign
5 Review shopping data. Either spend time on segmenting the shopping campaign or further keyword development Continue working on shopping campaign segmentation
6 In specific case, I might look at retargeting or DSA here. But most likely it will be more time on shopping or keyword development Finish segmentation of shopping campaign
7 Provide client with details on progress so far and an estimate of timescales for further improvements Go live. Report to client. Wait for data on what to do next
The time units don’t really mean anything

I am perhaps being unfair to the “fixed scope” method here - please let me know if you think I’m making it out to be a too much of a strawman.

But if your criticism is something like “I don’t wait until the end to launch everything and I do look at data gathered during the build process” then I think we are actually in agreement.