We need anticipate buyers using Search on G-Cloud to find candidate products

They also use Filters and may use Categories

With 8,000 SaaS products on the catalogue – it is essential to get this right or we don’t get found

This insight article tells you how

Buyers use search to find candidate products

The G-Cloud Buyers’ Guide gives detailed guidance on the process to follow to buy services on G-Cloud:

Write a list of requirements Work with someone who will use the service, buying specialists and technical experts to prepare a list of ‘must-haves’ and ‘wants’. These requirements will help you decide which search category, keywords and filters to use.

It also instructs users to keep an audit trail and record the keywords, categories and filters used because the process has to be fair, open and may be subject to forensic scrutiny. This is an important motivation to follow the guidance. The guidance then instructs:

Choose a category, then search for services using keywords and filters.

To be found it is necessary to have a strategy for each of these 3 components of the search, as they are applied simultaneously. With 8,000 SaaS products, buyers need these three tools to arrive at a manageable list of candidate solutions to assess. And the guidance on “What to do if you have too many results” – is to add filters!

Indiscriminate elimination by Filters and Categories

Filters and especially Categories are bad ways to eliminate candidate products from search. There are 251 unique Categories but a supplier is restricted to associating a maximum of 20 Categories with any product. A buyer for the police may search only in the Category “Legal and enforcement” on a fairly reasonable assumption of relevance. But this Category only contains 10% of the full catalogue and many popular horizontal applications (e.g. collaboration products) are not represented in the Category. So that choice would be prejudicial to finding the best solution. Similarly, faced with too many results, the guidance is to “add filters to refine your results.” This encourages over-specification by adding ‘nice-to-have’ features without consideration of the additional cost which the feature may imply (e.g. ISO28000 certification, which is only held by 1% of SaaS products).


The first and probably most important key to being found is Search. The fields which are indexed for search in G-Cloud 10 are:

  • Service summary (max 50 words)
  • Service features (max 10 words x 10 features)
  • Service benefits (max 10 words x 10 benefits)
  • Product name, supplier name and service ID# are also objects of search

Being found; selling your service to government” (2014, Ivanka Majic) draws attention to the important difference between searching for a product on Google and G-Cloud:

It’s not about being top of a list

“…I invite you to think less about being at the top and more about being the last one left after the elimination round. So when writing your description, keep asking yourself, “what requirements does my service meet?”

This is the opposite of conventional Search Engine Optimisation (SEO) practice and arises because the buyers are obliged to assess all of the services in the Search results and not stop after they have found a reasonable fit within their budget. This has significant implications for G-Cloud SEO. While the G-Cloud Suppliers’ Guidance states:

Clear content can help a buyer understand and, ultimately, choose your service. Any descriptions you include should be relevant, concise and written in plain English. Don’t include keywords. They make services harder to find and understand.

This is disingenuous. While the searchable fields should certainly be clear; a supplier must carefully consider the words that a buyer or analyst may use in searching for a solution, having recorded a set of requirements. In my experience ‘plain English’ means very different things depending on the department, the MOD and NHS use very different terms for software features which are identical and these can be different again from the terms used in the private sector. The Search engine used on G-Cloud is weak, it has a little fuzzy logic: a search on “invoice” and “invoicing” generates the same result, but “billing” generates a different result. As your product only needs to be found once (“It’s not about being top of a list”) – don’t repeat the same word or phrase in the searchable fields – but use synonyms that are likely to be employed by a public sector employee, this increases your chances of getting the one ‘hit’ needed. I have reviewed all the searchable content for a random sample of 100 suppliers on G-Cloud 9 which have not achieved any sales success (as at December 2017) and compared this data with the searchable content of the top 100 suppliers (by sales revenue) of the successful cohort. This has led to a number of recommendations which I set-out below; the simple rule is to think of the voice and words a real prospect for your product will use to find products that match their requirements.

 Do’s and don’ts for the searchable fields

  • Utilise your maximum word and line count, it is a computer looking for phrases and it doesn’t get bored
  • Adopt the right ‘voice’. You are addressing a search engine: “our service makes the others look slow” will only find you if someone is looking for a slow system
  • You need to cover the obvious, if you supply ERP, you need to use the term ERP
  • Many fine phrases are being invented in marketing to make your service sound unique: don’t use them until they are in common usage. You may like to call your system an XRM, but is that what a buyer will call it?
  • Do use words and phrases you know your target customers use that aren’t common in the commercial world
  • Consider separate catalogue entries for different audiences, NHS, Local Government and Police use very different language
  • At the moment you are restricted to 10 features, but if you list 2 services you can show a combined 20 features
  • If you have to repeat yourself in the 3 searchable sections try to use a different vocabulary each time
  • The guidance instructs you to not put in ‘additional search terms’ but it is essential to put in some! Ensure they are relevant, understandable and in plain English
  • Articulation of “benefits” is hard. Think in terms of what the head of department may have said which is going to be in the ears of the person typing search terms
  • We note that where there is a shortage of hard benefits, vendors list more features. As the search engine doesn’t differentiate between features and benefits, it is some sort of strategy, though getting the right benefit statement could be quite unique
  • Try reading blogs, public sector press, using words you hear in meetings and seminars to form a lexicon
  • Some services are so flexible and versatile they don’t want to rule themselves out from a universe of problems they can solve. But the actual gift is to rule yourself in. Don’t be vague, people don’t search for a product that can solve a wide range of business needs – they look for the one that solves their particular requirement
  • Don’t duplicate a feature and repeat it as a benefit. The search tool only needs to find a phrase once
  • Use a spell-checker


Categories, as explained above, are a blunt tool. I have raised the issue of inappropriateness for horizontal applications with CCS. The response is that buyers like them and so they will stay. Whatever strategy we adopt for Categories we need to start by evaluating which are the most suitable. Build a score-card and mark the Categories against a handful of criteria. How good is the fit to the sort of problem your product solves and how likely is a person in your target market to use the Category when they have that kind of need? If you have developed a marketing strategy with selected target markets, personas and identified the ‘aching needs’ where your product contributes greatest value – then use this marketing framework to score the Categories. Which Categories are most likely to be adopted by your buyer or analyst persona looking to solve the problem which your product best fits. Also look at where competitors and complementary role-model suppliers have made their selection (particularly those with sales success). If you have more than 20 high-scoring Categories, there is probably only one legitimate strategy: have more than one product listing and select different categories. If you only have one product on the present iteration of the framework, we will have to wait for the next iteration before adopting this approach. More than one product listing for essentially the same product also provides the flexibility of addressing the different language and needs of different departments and can provide a form of A/B testing.


Filters are susceptible to a different approach. Of the 17 sets of Filters, the approach to some must be to list additional products. For example, if your product can be supplied on Private or Public Cloud there is no alternative but to have a dual listing. This is a design weakness. It is possible to analyse the sales data and correlate this to Filter attributes. Ideally this should be done specifically for your product type. I have done the correlation for all products – providing a ‘general case’ of what Filter attributes appear to be favoured by buyers (this will be published in a future Insight article) and it does show that some attributes are highly important to winning sales. But the result will be far more reliable if you compare sales success and attributes for competing and complementary products. For example an air-traffic control system is not going to need to work on mobile devices, while a system for management of fly-tipping probably does. There are Filter attributes that are so ubiquitous that we might consider them to be industry standard or ‘best practice’, these are also best determined by reference to software type, but the ubiquitous attributes for the ‘general case’ will also be published in a future Insight article. In both cases identifying key attributes which your product does not embrace gives a clear marketing signal to either put achieving that attribute on the product roadmap or creating a tactical response to its absence. The difficulty being, if a buyer uses a filter to eliminate your product then the chances of them reading or hearing your tactical response are limited.

Exceptions explained

There are a minority of products with good sales which do not adopt the above recommendations and have very poor quality Search and Filter tactics.  Examining these in detail, they are exceptions at either end of a scale. On the one hand there are global, branded products that are household names, at the other extreme there are large highly specific public sector systems tending towards monopoly suppliers with a long legacy history. There are other anomalies but we can ignore these exceptions. If you want to build a sales and marketing strategy that becomes a logical and repeatable business process, maximising your opportunities to make sales, follow the success pathway and Get-found!