SEO Forecasting…The Holy Grail
I’m going to be completely upfront; this post isn’t going to provide you with a 100% guaranteed method for estimating your organic gains. So why continue reading? What this post is going to do is arm you with the knowledge as to why SEO forecasting is inherently difficult and the potential hurdles which await you. This knowledge will enable you to better navigate such difficulties and detail respective caveats to include alongside your forecast.
Forecast vs. target
First things first, what is a forecast? For me a forecast is an estimation of a future trend. A forecast can often be misinterpreted as a target. A target is an agreed goal which you agree to achieve (or at least make every effort to achieve). I am always keen to stress the key difference between the two. If you fail to effectively convey this early on, then it can land you with some very difficult conversations.
Why forecast SEO?
Whether you’re working agency or clientside, there is a high probability that at some point you will be asked to provide an understanding of the returns which are available from the work which you plan to complete.
If you are working agency side a forecast is a valuable tool which will commonly be produced to help push a new client over the line or win a supplementary piece of work.
Similarly, if you are working clientside a forecast will commonly be requested to be included as part of a business case.
If you are requesting budget / time to complete a specific piece of work, then in my opinion it’s a relatively fair request that you outline the benefit of said investment. Additionally, compiling a forecast acts as personal due diligence to ensure that your work is aligned with agreed KPIs.
Basic SEO forecasting method
The basic method of SEO forecasting typically follows the below method:
- Complete keyword research to identify target keywords
- Benchmark current performance (traffic, conversions, visibility)
- Apply projected uplift generated from targeted keyword ranking positions to current benchmark stats
Sounds fairly simple right…unfortunately not! If you base your SEO forecasting just on the above data, your predictions are comparable to those quickly jotted on the back of a cigarette packet.
SEO forecasting challenges
In order for an SEO forecast to achieve optimum accuracy the below challenges must be understood and where possible factored in to your forecasting model:
Keyword click data
Back in 2011 Google announced the introduction of encrypted searches for SSL Search. The effect of this was an almost total loss of keyword referral data within tracking solutions such as Google Analytics. This loss of data now makes it impossible to directly map entrances / events / conversions to specific keywords. As a result, work arounds must now be found through the usage of keyword click data available via Search Console, combined with estimated keyword metrics compiled from visibility data, projected CTRs and search volume.
In contrast to PPC, organic uplift is often a slow burn with permanent or at least semi permanent results rarely achieved immediately. Where PPC bids can instantaneously be increased to capture greater traffic / conversions (albeit at a significantly inflated cost), uplift from SEO activity often requires much greater investment of time and effort. As a result such trends are much more difficult to forecast accurately.
Impact of other channels
Other acquisition channels can notably impact organic trends. For example, a sudden and unexpected increase in PPC spend for brand keywords can result is likely to result in a sizeable reduction in organic traffic origination from brand attributed. Segmentation of brand / non-brand traffic is a crucial element of any SEO forecasting activity.
Above the line (ATL) activity
In addition to factoring in the impact of other online acquisition channels, it is important to ensure you don’t forget to factor in the potential impact of ATL activity. For example, at a simplistic level a TV campaign is likely to significantly uplift the number of organic visits originating from branded keywords.
Longtail keywords are those typically consisting of three, four or more keyword phrases. Individually such keywords typically hold significantly lower search volumes than shorter head terms. The challenge with longtail keywords is they are notorious difficult to research but cannot simply be ignored as while individually they hold much lower search volumes, collectively they are likely to deliver a sizeable chunk of the organic traffic to a website. Combined with this such keywords almost always have a much higher propensity to convert.
Very few industries experience a flat demand curve throughout the calendar year. As a result any accurate SEO forecast must also factor in such seasonality. Most businesses should be fairly knowledgeable on how seasonality affects them directly, so don’t be afraid to talk around this topic with them. Further seasonality data can be mined from Google Analytics, Google Search Console, Google Keyword Planner or Google Trends. The latter two sources are particularly useful for research seasonality for products / services which are new to the market or haven’t launched as yet. In such situations first party data from Google Analytics and Google Search Console are unlikely to be available.
Brand / non-brand
Organic traffic to a website typically consists of both brand and non-brand traffic. In order to accurate forecast future traffic you must segment and estimate the uplift on each independently.
Click Through Rate (CTR)
No accurate traffic projections can be generated unless without a realistic estimation of CTR. A generic industry standard organic CTR model just isn’t going to cut it. As a minimum your SEO forecast should utilise an industry specific CTR curve. However for optimum accuracy it is strongly recommended that you calculate your own CTR curve utilising Google Search Console query data or similar.
In order to produce accurate future projections, it is important to not only look internally, but also externally to review potential threats and maybe even opportunities. A potential threat could be generated by a new competitor entering the market while an opportunity could be generated as a result of a future planned event which could generate increased demand for a product or service.
It may sound contradictory but a final challenge to the creation of an accurate SEO forecast is the limited availability of certain data. Search volume data is a key example of this limited data. The vast majority of keyword research tools source such data directly from Google Keyword Planner via a variety of methods.
SEO forecasting will always be challenging, you will never be able to predict everything simply due to the extensive number of variables at play. Even if you were to overcome all of the above challenges, there are other factors which you simply cannot accurately account for. Future algorithm updates are a perfect example of this. Search engines rarely publish details of planned algorithm updates. As a result it is almost impossible to factor in the impact of such updates on organic visibility be that positive or negative. The possible exception being if you know a website is directly contravening best practice with a future drop in visibility inevitable.
While SEO forecasting may seem like an impossible challenge it is often an essential task which must be completed. It is key to understand the challenges involved in such forecasting and ensure you invest effort to overcome these wherever possible. Finally when delivering a forecast be sure to effectively communicate respective caveats.