Improve productivity, visibility and agility with
TestBoard is ready to roll out soon to a small, select group of beta testers.
We will never ever send you spam, nor will we ever sell or rent your details to anybody else.
Why we built it
Through our many and varied consulting projects, we have worked in different organisational situations; with client’s in-house teams, with agency teams and directly for C-suites. In all cases, we have sought to deliver, indeed our clients have come to expect, a rapid cadence of incremental gains from our model optimisation.
To deliver rapid iteration, we have to continually push ourselves to improve collaboration between our own engineers, mathematicians and strategists. Not to mention, between us and our clients and in turn, their colleagues too.
We have sought to maximise collaboration, transparency of process and accordingly to drive productivity into our service. We always seek more measurement and more accountability from our own colleagues and our clients.
As clients grow with us, so we see them begin to think differently about data-driven decisions and the manner in which they brief us with questions, hypotheses and challenges.
We totally believe in the agile method of team management and think that it is the way to speed up incremental gains in optimising marketing and customer value budgets. So, whilst we admire and have tested many of the brilliant software scrum-boards that are available, we felt that our data science in business intelligence, marketing and business optimisation deserved it’s own tailored method to acknowledge the difference between these disciplines and to enable us to deliver consistent and value-creating iterations.
We built TestBoard for ourselves, but now we want to share it – put it out there, see what you guys have to say about it and see whether you find it useful. So far, our clients really like it and the feedback has been good in closed beta, but we have wide-open minds to hearing feedback and developing features for everyone to use.
We think TestBoard is for any part of the business that wishes to optimise it’s activity working with data science colleagues; a tool to organise your testing; to collaborate with your colleagues and agency partners on challenges, hypotheses and good old-fashioned, searching questions. It’s a tool to help prioritise which tests are likely to generate the highest returns soonest and to help you speed up your testing habit, so you learn more, faster and optimise ahead of your competitors who are also improving their data-driven testing habits.
TestBoard has already been used to deliver significant business savings and reinvestment by retailers, utilities, charities, FMCGs……. so we would go so far as to say TestBoard is guaranteed to increase your cadence and hence help you optimise your budget faster.
Just like any scrum meeting, it starts with ‘bright ideas’, challenges, questions etc. all of which need gathering up with energy, organised or ticketed, to harness the collective intelligence of the team. By ‘ticketing’ assigning a question or challenge it’s own ‘card’ which has our current agile methodology embedded. Here’s how the method works……
[the below paragraphs must be associated with up-close screenshots & highlights of relevant features of the card’s inbuilt methodology]
Ask the Questions….
Everyone in the business, marketing team, the company’s marketing agency vendors and the Business Intelligence team are being asked questions, seeking insight into how to optimise the business’s objectives. These questions, large and small are typically not logged centrally for the purposes of dealing with them in a logical, value-driven order, to test and learn to an agreed programme.
TestBoard is for logging the questions and hypotheses about marketing, advertising, promotion and audience that all businesses have. It will have common business and marketing challenges pre-loaded by sector, because we have worked on them time and again, but all businesses have different questions, so you will want to keep logging your own refined questions.
Do the Analysis….
Once logged, your data scientist, business analyst, team member or agency staffer has a chance to examine the data and calculate preliminary valuations. Some businesses describe this as scenario planning, to value how different variables could affect different investment inputs.
Prioritise by Valuation
With a log of questions posed, some simple initial valuations on the potential Return on Investment (ROI) for optimising activities, helps to prioritise the order of testing to learn in all channels.
Test, Learn, Repeat, faster
Once you’ve valued and prioritised, done the statistical analysis, you can adjust your testing and learning schedule more frequently to optimise more quickly.
Now, like many of the fastest-moving, younger, disruptive businesses in many industries, you have become a data-driven marketing team by acquiring the habit of using data to prioritise your ‘test and learn’ schedule.
The current closed beta is still in a pretty basic ‘ticketing’ functionality state. Here’s a little insight into what we’re ‘testing ugly’ for use in our professional services …
On-demand analysts and resources
If you already have an optimum number of people in your team, great. If not, or you believe in the wisdom of experts, crowds, second opinions, or combining all three, you can assign trusted resources single card problems to generate insights.
Pre-loaded model library
We have an expanding library of models that answer commonly posed challenges. It can be helpful to begin with ‘oven-ready’ models that need tailoring to your challenge along with the common APIs required to populate data from BI, ETL and consumer data sources.
Subscription on public data sources
Often our models perform better with fresh inputs from new data sources that our clients subscribe to. Many of these are free. We will pre-load these, so that you can overlay your internal data against public dataset trends.
Have a bot do stats work for you….train the AI
There are common statistical analyses that take time to do manually; so we will embed IBM’s Watson, Google’s TensorFlow, OpenAI and our own customer-sensitive bot, Sherlock. We are building this semi-automatic statistical analysis to begin by running some standard statistical analysis on all your data sources looking for correlations between inputs and outputs, seeking patterns to match and produce insight without you, your colleagues or agency staffers spending time on the valuation. They don’t don’t understand the nuances of your audience, but they do the common basics quickly and effectively. The more you train them, the more they learn how to do it better next time.