Data Analytics in procurement: opportunity or risk

welcome my name is John Broome I'm the chair of the contracts and procurement sig what are we going to do today well it's just a little bit of scene setting in a very brief introduction to our speaker and then we will crack on with the main event a little bit of sitting setting Martin approached to do this webinar and we discussed it our committee meeting and I have to say we were very divided on it and the reason was that there was a concern amongst the more conservative should say more legalistic based as well people that we would inadvertently be encouraging people to do stuff which is illegal so what we decided to do was go back to Martin and have a discussion with them raise our concerns and fair play – Martin Martin said you're absolutely right to have those concerns they need to be discussed they need to people need to be aware of it so they can either decide not to participate or go forwards with this or if they are are they going forward with it with their eyes wide open so with that I'm gonna just very briefly introduce Martin so Martin is the founder London and Bristol data analytics meet ups which is a community over 2,200 people in the interest of libri interest in levering the plume of data that is emitted from projects following the APM a registered project professionally chartered engineer and CEO of and founder of projecting success consultancy which specializes in data analytics thank you John really appreciate that thanks for the opportunity today as well in terms of data analytics and bids I think it's a exciting times actually I think there's going to be a big transformation coming across the industry in data analytics working on projects and that's something that I've been focusing on for the past two or three years now and so just like to weed you through a bit of my background and explain a bit of context the first slide is about my sort of background so since we mentioned already I'm a fan of my childhood professional and a charting engine have worked in loads of different sectors and more recently I'll be working in oil and gas and construction but I worked in aerospace defense nuclear and big ICT projects as well I've been a project manager which I led a billion-dollar job I've also been a portfolio lead when it appear mower on a multi-billion job as well so I'd quite a lot of experience working a supply side and bio side and I've been into various jobs as well so I've seen this from both angles and for me I think there's a great opportunity now to start to transform the way we start a big projects in the future and we start to analyze those and the way that we start to engage as suppliers in that process and it's potentially transformational so let me give you some sort of overview in terms of what project data analysis is all about first of all so there's some terminology that's floating around so artificial intelligence is the term it sits right at the top underneath that is machine learning and what's really important about machine learning it's about learning from experienced so people talk about lessons learned and lessons learned is something that's been going on for 40 50 years and it's not that effective if we use slightly different language which does the same thing so instead of lessons learn we start to learn from experience we commence not to use all of these on funds techniques and these are bonus techniques that means that we can start to leverage all of that experience from bids from project delivery from risks etc and start to join all of that together and get some insights which bank and kriti achieve in science as well in terms of deep low and then it sits underneath machine learning and it's in your network which is automatically created by the computer and it's tough to look at those features across that network which differentiate and which starts to give some predictive capabilities so why is this taking off at the moment so it's taken off the various reasons and a big one is data there's just so much data out there at the moment and there's a good quote there and of IBM's quote which is in 2016 90% of the world's data in terms of all data ever created had been created in the previous two years and that's still continuing to grow now we've also got clouds and that is cloud storage and cloud compute and we also put the algorithms in the cloud as well and we can access those through things like Python so a lot of those arguments are actually free and they've been created through PhDs and you can also go and download them from various facilities such as I've been Watson you can do two tensorflow and Amazon Web Services etc so that's a quick overview about what data analysis is all about so in terms of the use cases so in project management it's worth just whizzing through some of these use cases so what we've got in terms of risk management we tend to look at risk as a once through process and a risk management is used for project delivery after project delivery all of those which tend to get thrown away or archived and we don't use them again if we start to join them together we start to look at the wrist lifecycle and see which of those risks turned into issues and then what lessons we can take formats such as that's when it starts to give you some predictive capabilities we think about stakeholder management's we look at things that are moment with things like power versus influence grids and instead if we look at that as a connected graph and we can start to get a lot more insights from the data and that graph there on the right hand side is something which came out of the game with phones and somebody took a script and they read this script through a computer and they look at those relationships between key characters and the power of those relationships in terms of the number of communications and start to map it out and you can press a play button on that you can see the evolution of that stakeholder grid over time so the thing on the left hand side is what we use it at the moment which has a bit 1980s 1990s and I think on the right hand side is where we could start to move to if we start to use a bunch Plato analytics and that was really exciting if you also look at Serge so we can take the corpus of all schedules and were produced in in Tyrion and some organizations start to do this we can extract some logic in terms of work great now a structure working and looking to it in terms of deep learning and find out which of those have parts of the schedule are most subjective to variance and we can link it with weather models and things like that we could start to predict then what the out turn of those projects are going to be and that's not based around as simple things like Monte Carlo simulation we can now start to do it much more effectively so that's a use case in terms of project management and now I'd like to drill down into the use case associated with analytics so this presentation I would warn you it's a bit more mighty and there's some people who think that this is not what we should be doing we should leave bids to market forces and not start to analyze the detail whereas I believe cuz I'm an advocate of project data lytics is that once you start to get the insides out of this you can start to really shape things and start to make a lot more intelligent decisions on working with the supply base and as a bidder also bidding into it so we can then start to really shape the future of procurement so for me it's all down to perspectives and different people who's trying to stop this and I can understand why I know some people who's actively trying to push it forward and I'd say try not to look at it through your own perspective try and take a step back and think of it in terms of the macro picture and the macro economics a bit and for me it's not a case of of should we do it because we can do it already and I'll start to demonstrate some of that and people will start to apply this because we're working with people at the moment and we see that this is all possible today I think the question for this community is what do we do to shape it into influencer so in terms of those use cases we've got various things which are operating in tension in terms of public sector they're looking at using this to encourage of maximize competition scenario modeling improving quality of bids etc but they're also concerned about preventing precedent so if we start to do this somewhere then once it's been set then they're worried that it's going to be set across various other bids in the future it also comes on at people a game it a reputational damage if we find from this analytics that some decisions be made which shouldn't have been made and they'll be very suppose we can now start to pull the outliers out of them and a third thing about tension is the link back to the industrial strategy so there's an industrial strategy which is about artificial intelligence sector deal and it's trying to push this sort of technology and it's trying to say if we go down in artificial intelligence future we can now start to drive a lot of exports and we become really effective as a country so we've got all those things are working in tension and then we've got even more tension which is with the private sector so if we start to think about this so we can start to scan these opportunities in real time across all these various procurement portals and we can download the text from them and we can put some text analytics over the top of them and start to work out based upon all the bins that it's been bid before and which ones have been worn which ones are being lost etc when you've got the greatest probability of win so you can start to rack and stack them in real time just from sort of web scraping data and that means then you can start to work your pipeline up but it's not post a personal bias influence it's from the data of performance of what's gone before and I'm not saying that should be the answer I'm saying that should influence your decision-making it can start to inform what you're bidding you don't bid it starts to inform scoring mechanisms as well so you can go back to the public sector and you can say the public sector for instance that you've got these big scores in here and because you're waiting x-parameter really really highly it means you only got one really strong competitor and that means you've got a dominant competitor and they're gonna just clean up on them market and they can name their own price we start to change those scoring mechanisms it can start to level off the playing field and increase competition so he could identify areas for improvement shape future strategy in terms of marking gauge Minh strategy and you can also look at client loyalty as well I think that's quite interesting so which clients tend to work with the same suppliers all the time or which ones just go for the most economically advantageous tender a collaboration as well so you can work at which partners are the best view and you can identify bids which have got an outlier so where you think the sort of bid scoring's works against you I think that's quite interesting as well and scenario modelling so you can start to say if we change this part of the bid then what's the impacts and if we pull this through from another bid then what's the impact so all of those use cases are working in tension and I except I depended upon where your Sat you've probably got some perspective on it so in terms of this presentation today I'm going to be talking about construction bigger I've got a lot of data from construction and and it's quite easy to talk about we can do it in other sectors as well but I think this is a good use case just to get the conversation flowing so in terms of construction let's just look at the backdrop and the economics of construction so there's all these problems with a chameleon in inter servic cetera and a big part of coolin which was pulled out in the Parliament report is about low balling bids under bidding etc and we can now start to detect some of that so we can look at that in the overall data set that's very difficult to do unless you drill it into the data what we can also do is to look at the contractors in the industry and the top 10 so this is from a building code UK from July last year is the top 10 contractors I've got a margin of nought point 5 percent so if we're looking at that we need to be working with the construction sector is to make it profitable so we don't get more failures because when we get frail is government's got to bail them out people lose their pensions and there's lots of people haven't worked so the macro economics of that really see difficult so we need to be working with the industry to make it commercially viable so we've got some healthy competition etc so we need to think differently about this problem and there's another slide there so that was very recently so it's 13,000 construction firms in significant financial distress so when we look at is through a public sector lens we'll think that's one competitions we want the most competitors involved those competitors then going right to bid bid cost some two hundred thousand pounds to write the bid if you've got a lot of people bidding it and they've got no chance of winning it you're going to be contributing to those 13,000 firms which are suffering financial distress so we need to think about the macro economics of it so what we did is we pulled a load of data as we've got it from Ted dating a contract find a data source and subscription services we got 70,000 contracts and some of those contracts aren't construction contracts and some are so we then filtered them what we also did is we put in a lot of Freedom of Information requests and for me that's a unsustainable approach I think some of this data should be in the public domain so we don't need to keep asking for it and I'll explain why so in terms use that for my requests I just like to build this picture up field so we've been saying for instance they could you share the bid scores on a certain competition and this is the scores so if I say a maths test I'm saying I want to know I mean what my score was in long division multiplication I didn't know and taken away I'm not asking for the answers to the questions because that is commercially sensitive I just want to know what the scores were so what we can find out and this is already available at the moment and there's loads of precedent on it so I can know of who won a competition and that's in the public domain that tends to be in Ted you want to know who competed then that's available what the winners score was versus the rest of the scores and that's available so you can start to sort of mind some of that data the composition of the winning scores have been getting that information from the Freedom of Information requests largely without context there's some people that who said that they don't want to share that data and we come back and appealed that and we've won most of those Appeals the composition of all the scores as well so not just the winning scores but all scores would be winning that and the only one that we've not won is who scored what so we know that the winning score would score six in math attending a subtraction nine in multiplication what we don't know is what the rest of those organizations scored and if we don't know what the rest of the organization scored we can't do a relative comparison from bid to bid we can't start to understand trends so I've taken this to the Information Commissioner's Office and the ruled in my favor against the Cabinet Office and the Cabinet Office have now appealed to tribunal service now Co ruled against me on a another one of the cases were identical it was looked at by different case officer so I've appealed that one to the tribunal service so this is now going through and very timely I got this from the postman knocking on the door this morning which is is the case packs so this is now going to tribunal I've asked for a hearing because I think this is really really important that we start to test what goes in the public domain with these things because I think this is really important and let me explain why it's important and what the process is so with big data it terms a worked example we take all of this data from Ted and various other places and we clean it and there's lots of things in there lots of typos lots of mistakes in there you try and pull those out we then associate the data so we say these things are talking about the same bid so we've got it from different data sources and we aggregate it and I'll link it together we then apply text analytics and segmentation to it so we can say for instance it's talking about the school is talking about hospitals in clients so if it's high-res England tend to know that it's about loads if it's round abouts then we know it's it's a certain sort of construction project and then we can play machine learning in say May upon all that information all the big scores who's won the cab data etc we can now start to forecast who's won a bid etc so we pull out of all that data these things called features and those are the things that use as human being normally to a form of view on what the factors are which are going to differentiate a bit and when a machine does is pulling out thousands of those features and that might be things like is it roundabout as a school is it a certain location what's the value with the bid as well cups of small bids tend to go to these sort of organizations and the bids above hundred million go to some organizations so it picks up all of those parameters and it starts to put weightings on them and says in these situations these features are more influential than these other features so we can then start to find out who's likely to bid who's going to win and do the win by a nose or they're going to win by a Furlan who's in the top three the top five on what the probabilities of each is a success rates of each of those suppliers based upon a number bids that put in and what the client loyalty is and variance in performance across various clients so you might find some bias in certain clients against certain suppliers or it might be a positive sort of bias which works in the favor being some supplies so we can start to drill into that and it's all in that dataset so this is when it starts to get really really cool so we could predict bid outcomes to an average I could see of 30 percent and you think well that's a one in three chance or gaining right now that's based upon an overall data set so if you've only got a few data points it means that this data analytics is not that good it doesn't perform that's just data analytics for you it's just having a spread of data it's down to the quality of the data and with some of these at the moment we've not got a big feedback scores we've got some big feedback scores but not that many in the overall 10,000 contracts from the construction industry so once we start to plug that in then these performance levels will really start to shoot up so in that work well for instance we've been forecasting the accuracy of Network Rail bids to fifty percent so who's going to win a bid we've been getting it right fifty percent at the time if we look at North Tyneside for instance because there's a framework there they've got four contractors in it and there's been twenty feed contracts awarded we've got rad that rides 21 percent of the time and these are going to need to be refined all the time and we've been working it through in terms of a Minimum Viable Product so this is our first go with this proof of principle just to see if it's possible to stimulate this sort of debate and you'll see there's a lot of variance there from from authority to Authority and we're now starting to understand each of the parameters that start to influence that I think it's a really fascinating science so what's the sort of thing that we get out of it so what we can start to find out is the five most likely contractors for a certain opportunity and in that you want to understand those five contractors got an in-court chance or as one got a dominant chance so you'll see there with bound construction for instance they've got an overriding probability of winning that's a certain bid and if you were balfour beatty then your chance of winning it will be two percent and that's going to affect your big know good decision there's loads of other analytics which we can start to drill into them so in terms of these sectors as well you can start to work out which sectors you're performing her best in you can look at supplier churn so the certain clients keep on using the same suppliers all the time and joint ventures so which joint venture partner tends to perform the most are the best etc now we can then get into the big score so this is where it starts to get really interesting and this is is where we start to focus some of our attention and this is where the Serb tribunals going as well so we've raised that for a while request against 42 frameworks in around 200 underlying contracts and we've got a lot of scores from them so we can then start to do all this analysis about where your competitors start to outperform you KPI analysis in terms of a greater spread of scores compared to average so what's the overall variance and we can start to look for certain question tags because we've got a taxonomy of this one question types where you fit on the overall distribution you top quartile bottom quartile etc and it is a big distribution it means that some of your suppliers are probably bidding and they've not got that much sort of covered Lituya of winning and I think that in terms of clients I think you've got responsibility to help those suppliers to improve and I think some of this starts the feedback and I've seen from this scatterplot here in terms of smaller organizations they've got a sort of lower average score overall so that's the weighted score against the standard deviation of the weighted scores and the larger organizations tend to be clustering around the bigger scores and more consistent scores as well so they go to next slide this really starts to explain it as well so the big companies which have got the more consistent scores are not more successful and the companies who are smaller I've got a lot more variation in their scores because they're not that consistent because they've not got this off to a tee yet they're probably not got full time big teams etc and they're spending a lot of money on bids and they're not winning them so we need to look into that what we can also do is we can look at from framework to framework what the consistency in those bid scores are so they look it's a Robert MacAlpine there and we've taken a sample of those two frameworks their scores are pretty much identical from bid to bid if we take Obama construction in one case and gods a 61 percent of their overall score on one framework and 40 percent on the other framework and it's a very very similar framework and you cannot argue so what the difference in those two organization may mean those two clients why is it scored so differently and then taking care there as a prime example in one case their score is probably almost doubled so the black one at the top for your build is almost double the one at the bottom which is for the school's framework and if they're bidding things and we can start to drill into the underpinning detail of those scores we can now start to state why are we getting almost 50% of the score in one framework versus another framework and we can then start to drill into it what we can also do is start to do some box and whisker plots so just to explain what these are these are about quartiles basically so you can look at for a specific frameworks and we've taken a sample and the frameworks here you can start to look at the variance in scores across framework so we could look at have been for instance of Building Information modeling we'll see that there's a clustering of those scores and it's quite a small variation if we take things like health and safety and risk management for instance then from framework to framework this is the average scores tend to be very very different and the minimum score those for 2% the maximum score is 75% so that's worth coming into and to understand a bit further so that's at an aggregate level looking at each of those frameworks and working out a framework level what is the composition of those scores on average if we then start to drill down and we look at a random company we've picked how is Morgan symbol you'll see there is that some of those scores have got a large variance so the one there on business resilience and continuity there's a variance which goes from 50 percent to 92 percent across three frameworks and that for me is really surprising because if I write in a bid that content of a grid it's probably going to be quite similar so why does it vary by 42% that means looking into an acting drill deeper into that and I pulled some other ones out as well so health and safety tends to be a very similar proposal from bid to bid on health and safety that's very green by 30% and scores why is that we need to drill into that and start to understand it and we could do this against each of these line items one by one and we can start to understand why we've got a variance so that's for my supplies perspective and from a client perspective we can start to say why we get this variance in our supply base what do we need to be doing so we can lift up the capability because what we actually want is the best possible bid it's the most competitive bid what can we do to help them and to improve all of these bids in the future so in terms of doing this what we're trying to do is they get some consistency between these frame weights as well because if you're bidding one it's a bit of a lottery if you bid in the second one and you're getting 50% of the score for the first one for no transparent and visible reason it starts to frustrate you now we can even work with suppliers as well because we need to close the gap from this front honor if somebody's really dominating the market but we can bring it in so can try it another competition what we also might want to do is start to drive a more collegiate either instead of doing everything in an hour and hour meet their competitions we might say let's work towards more collaborative frameworks and money's some of this differently I think there's also a case as well if there's various frameworks and up to 60 common structural framework something something like that do we want to start to mark some of these things which have got commonality from bid to bid like health and safety and BIM these other things do we want to mark them centrally by some divine cabinet office so we get consistency on scoring across all of those bids I mean that's a question and a bit of challenge where we need to be asking some detailed questions and I keep saying as well this is about the macro lens it's not about individuals it's not around departments it's not about your reputation or damage we need to look at the overall picture so now to the wrist and I think these wrists are fairly significant there's miss of collusion so if we're sharing data from a company to company and that's not shared with everybody that is collusion and you'll go to jail for it so we need to be really really careful about doing that and I'm sort of been talking to lawyers and getting some legal advice on what's a lot of what's not allowable if we start to change the scoring mechanisms as well so if you do it amid competition that's obviously not allowed if you do it in advance of the competition that what's the implications of that I think we need to think through them what's the implications of using legacy data as well so if you now start to say right I've got some scores from previous bid on health and safety and BIM and I'm going to use that again on this bid then what's indications of readings in some of those scores and there's a perception of blacklisting as well so if you've got a load of information on people as a perception that you're gonna use that for the wrong reasons so if you start to collect this data start to use these sort of capabilities you really need to understand what the scope is of your use case for using it if you use it for the wrong reasons you will go to jail I hope you comment miss I've got a few slides on this actually so I'll touch on that in a minute and in terms of data volumes as well so I just like to point out that this doesn't work very well with very very small data volumes if you're building one school every blue moon in the Hebrides it's not gonna work if you're doing it and you're putting Railtrack down at time after time after time you've got lots of data on it and it works really really well in terms of reserved the results as well you gotta be careful because there's bias in the previous results so it's biased against modular construction for instance because modular construction is not reading the previous results so you need to take account of that some of these suppliers have gone like cooling was not in there anymore and some strategies will change as well so it could be a market entrance or it could be somebody was pulling out of a certain sector so you need to be careful about using all of that data without putting some intelligent insight over the top what we're also looking at in terms of these scenarios is collapse scenario so we could collapse competition and these dominant suppliers Double Down crust they've now got the data and they can really start to drive things forward if he can now start to understand under bidding as well you can start for stands a nugatory bid costs you can collapse competition can also start to get competition to thrive because it's less of a lottery you've got more transparency so reduction in in bidders but we get close of a cluster in and we can start to shape that score in as well and we can also transform it so a third scenario is we can start to transform it and I think if we can start to look at some of these factors one by one we can really drive some transformational change that's positive the clients and suppliers and the entire sector so in terms of impacts on procurement professionals I think we need to get a greater understanding of stats and data science all this transparency drives accountability gives us this outside view we can identify variants and bias and our analysis I think we should embrace that not push it away I think this is really disruptive so we need to start to understand emerging case law which will start to come out when he took in requirements as well and the scoring where we can encourage competition and if it's not a real competition you've got 20 people bidding for it and it's costing society a lot of money to do that so what can we do differently and there's a potential as well for increased automation we can take this a lot further so what's the overarching pictures for me is about the civil service code as well I get a lot of people in the public sector who's very young to this because they think well it means we've got more challenges and stuff if you need your civil service code this is what it's all about it's about integrity and partiality it's ignoring inconvenient facts and things like that if you're unjustifiably favoring one party against another one it's all about your civil service code so what we're talking about aligns fully with advanced data analytics the problem is at the moment we're putting a lot of this into freedom InformationWeek worse and it's really hard so we can now start to look at these concepts of data trust and data trust we're cracking on it pace we dated trust at the moment and working with the oil and gas technology center and the only gas authority to set one up on a gas sector and that's a regulator and that's the share project management become a data and we're also about to do something very similar with some Robert mcalpine in the construction sector as well so if we can do this with big data and the bid analytics that goes with it and manage that by the trustees and make sure that we use it for the right reasons we can put a safe environment in place this is not about sharing data it's about securely pooling data against specific use cases we can then start to really drive this forward and get a wealth of insights out there it can transform quick human so in some way people say – this is all pie-in-the-sky recurrently we can do it it's all possible today people are start to use it it's start to engage in it so what do we do to shape it and once we start to shape it that's influencing for both parties it's going to lead to a lot more challenges in terms of probably procurement decisions but that's a tellurian in due course I think some those legal issues of those risks are went through our gray area that probably needs to be your bottom doubt and it's gonna transform public procurement I think all procurement and people say to me so what does the future look like and I say I can't predict what that future looks like because it's emergent what we need to do is work as a community to engage with it and to react accordingly and make sure that we do it safely and we do it with respect and we do it anticipating the potential use of some of these tool sets as well so John that's the main talk completed as the screen suggests we're using many to me so I put that thing up there there are loads of questions on there either go to mentor meter comm or download the app and answer the question so that's another reminder especially now that Martin has given his overview the questions are starting to come in one through fill it and he says how you at uhm how much of the scouring bones to put down to the competence of the buyers and how much do you put down a little level of training that given to note how to evaluate this eg how often this task is going delegated to a project sponsor who has not done it before once we start to drill into this data and this is very top level data at the moment from what you've seen once we start to drill into it we can start to get that understanding so I've funded most of this work as a Minimum Viable Product just to get a conversation going right and to say is this what we want to do now if we do want to do it then let's work together to start to answer some of these questions and say or is this telling us what is this tunnel is about competence where do we need more training and if there's one authority who's scoring completely differently to other authorities then an intervention is probably required because it then becomes more of a lottery if you bidding into it and I think that if you are spending two hundred thousand pounds on a bid you need to understand your probability of wind you know in broad terms it's not going to be exact because it's a probability but is you need to know are you in the race or you a donkey who's in the back of the horse race you know that's where you need to start to understand so I think it's an emergent process John okay got another one front nerve says have you looked at the relationships between successful bids stroke bidders and actual contract performance each Association successful project a delivery now I'd like to add I add in another one into that in that have you looked at how at the point of selection you can't use past performance in public procurement but you can a pqq stage as indicated on that final bid list what are the legal issues there so first of all nerves question have you looked at the relationship between successful bid stroke bid as an actual contract performance and the answer is not yet but in terms of that datasets there's a lot of that data in the public domain so there's things like the major projects report from the infrastructure projects authority so we can start getting from there and you can also put food and information requests in and there's some frame weights as well where they publish those results all it needs to do is to go into something like a graph database and you can correlate the two and get the analytics out of it so it's possible it's just the willpower to make it happen and the effort involved in that so I think that's definitely possible more I think we can do that it's just a case of when and I think as well if we want to do this as community then let's get a backlog of all these things that we want to do so this is like an agile process so we've got this Minimum Viable Product what do we do now to build it out so we can all start to engage in it and start some of these conversations and transform procurement what's the second question John second question was what would be the legal issues in using all this hard data at peak a huge stage to say well you answer the questionnaire some lines and it's very ill formed you answer the question they look good but actually when we look at your data you deliver poorly or the other way around for that matter so I'm not allowed John and I think that you're sig probably understands that a lot better than I do but I just think that this capability is going to test the law in various situations and and I think we just need to examine each and every component of it we can't go into this blinding just doing recklessly we really need to think it through I think it needs two working groups on it and try and work through what the implications are yeah okay so the answer is I don't know but I know it matters something like that yeah this is from Katrina does your model learn from its mistakes and correct accordingly over time to improve its rules and accuracy yeah so it's a full machine learning model at the back of this so that slides which are presented before with Network Rail on it and I think was Northeast Council on there etc that's learning all the time so every time we get some information in and it gets it right or wrong in terms of the bid for it's always learning from it and that's just using standards and machine learning algorithms okay you've got a question from Samuel he's having a hard time seeing how this applies to the private sector where freedom of information is not clearly available what's your views on that so for me it's the same challenge I think so I can get this data at the moment from the public sector because I've got a tool to do it but if you do it with the private sector if you're putting in a bid and it's costing you hundreds of thousands of pounds to do it I believe that we should go in a conversation with clients to say can you give me some feedback on that and can we openly share it if we can openly share it then let's start to learn from it so we get better bids in the future if you start to get lots of people competed on stuff and their bids are not that great it's costing you effort to assess those bids and it's not a real competition because they're an outlier and they're never going to win it and it's costing them a lot of money because they're working on false hopes and false promises so I think once we start to do this if we start off in the public sector and we demonstrate we can do it maturely and not in some sort of confrontational type of coach I think we'll get there so I think this is a cultural change and I think this transformation will shape that cultural change okay this next question leads on from that and it's from Jason and he says I understand the perception of collusion is important and you need to address this how does collusion interact with data trust if data trusts are only signed up to by a select few organizations so the concept of a data trust right it's really really important to understand this if you google open data Institute and data trust they published some reports on this about two or three weeks ago and I've been building off the back of those so what they're saying is with a data trust it needs to be independently stewarded so if you put your data in you can't just go and look at other people's data all right so the steward would look at the risks associated with releasing that data to another data provider or to a third party and this data steward basically would go back to the trustees and say and one of three things I can use this data it's fairly low risk I don't want to use this data because the TV progress or this collusion miss or it's debatable if it's debatable it goes back to the trustees and the trustees are representatives of those data providers all right so those data providers would say in this case we've done a risk assessment we believe this is borderline occlusion where they need to take some legal advice on it or we don't provide it so it's not a matter of the fact that if you in the club you get some privileged data what we're trying to say is everybody can access that data in theory if they've got an appropriate use case to do so what we do for instance is to give that data to a third party organization it might be an innovator who's coming out and saying if I can process all of that data i can give you insights now on the best way of getting bid performance off by 10% for everybody and if somebody's got an algorithm to do that and that's great and that would work for bidders and it works for clients as well so it starts to tell you all of that innovation right okay my last question which is just a very simple yes-or-no Martic is presuming the steward of those data trust cleans up data etc and makes decisions on that yes we didn't quite right so what we've got at the moment is we're trying to drive forward this agenda about project data analysis and its transformational capability so we're trying to pull together a project management people portfolio and program management people about my new project mind as it might be ever become with professionals bidder is whatever which one important in the same room as a data analytics people date of science it's free to join as a free community to attend and it's a magic source by bringing those organizations together in too much time we've got this thing called a project hack and what we're trying to do there we go a hundred people in the room and it's at the Microsoft reactor in London and we try and give people some hands-on experience of working with project data analytics and start to explore some use cases with some real data and if you've got day tune you like to come along you'd like to present some challenges and please get in touch with me and we'll try and make that work but this is a really transformational capability and I think it's in a lot of people's interest to start to understand it and start to understand its performance capabilities and we've got these national events so it turns a London meet up it was founded back in December 2017 and it's now over 2600 members the Bristol events just taken off recently and the Manchester events just taking off recently as well but I think we're just about to pass over all we just want to pass 3,000 members which is a great achievement and what we're trying to do is to get this critical mass where we all start to talk about we all start to talk about data we start to pull this day to day to trust it then gives us all the capability to really shape the future or project delivery and it's really transformational a lot of people's jobs are going to change as a consequence for this and if you want to get in touch with me then there's my LinkedIn and if you want to sort of find out about the project date on it it's community then there's a LinkedIn a QR code there as well and that posts on events and if you find something interesting as well then we post through LinkedIn as well so it keeps you plugged in with what's going on there in the community say that majority are from the private sector people are about double that actually listening into the webinar so we can see where people come from the sector's most of them a civil engineering construction or obviously I need to refine my against AI feedback though the samples that's not big enough on what those others are and we can have a look at what people do most of them are noticed they referred 30% hands-on project managers only 4% procurement and 7% because commercial and only 9% actually involved in bits which I think this is coming to bits near you so that needs to be got up most people are on the bio side 62% almost thirds and then when we started to look at questions the reactions we've got so provoking well that one over there which is a little bit negative most of them are positive though it has to be said though Manley's I think that this one over there there's been a lot of birds but I've enjoyed seeing habit he how it can be used I think Martin's presentation has brought that alive to the potential of it this one here a real negative risk of corruption illegal change and it goes back to our reasons for running this with I think it's going to happen in some form or the other you can either knife decide not to do it or if you're going to do it or explore it you've got to absolutely do it with your eyes wide open so if we now look at some of this stuff this is having heard the presentation people feel that most people feel they can give some form of an informed opinion some people about third size still don't press I'm still processing it more reflective people and others I don't understand it most people or 40% think we should help to help to shape it and a few thing we should actually try and block it and some should just keep passive eye on it and the impact are I'm surprised personally my advice people think that it will be quite significant to transformational now that implies that actually most people think that in some ways it will happen and logical conclusions is that if in some way it is going to happen then you need to shape it and take part in some way in there oh 42% I want to act no more while 35% want to get actually get involved and any unresolved issues where you can see this so in this chart the size of the words indicate how many people mentioned the word if you like so from that we can see and it's evolving as people at it so from that we can see that legal and data and legality issues and legal issues and legality are in there are the big ones and the other one stands out a little bit is ethics and I think that's relates to the day to trust and the use of those so as was in an ideal world data or legality follow lithics then we need to sort out the ethics first before we do look at the legality of it and obviously in the real world it's evolving so in terms of the last one I think the challenges are showstoppers so 5% think they're showing the stoppers majority not by much think they're significant but surmountable and 45% think that all part and parcel of imagination unmanageable 39 think the benefits outweigh the downsides so which implies that well I won't say it applies that five John Goffe but there might be well be some people in the middle as well and in terms of that last one the concept of the data trust I think generally majority or 37% are take that it's it's something which will resolve some of the concerns but at the same time people think this it's going to take a long time to mobilize and let's look down there almost 1.9 percent think that it's not gonna so close out your thoughts your summaries I think that my summary is of Marty's presentation is that I can now both in a project level and bidding have a better much better idea of how it would work I think there is some real good potential both on the procurement side and on the bidding side to transform might be strong word but tweak is definitely an understatement significantly modified personally how people go about procuring and especially in the supply side deciding to bid but I think that there are legal issues which need to be worked through and possibly challenged so without a reminder to draw your event noticed two forthcoming events one is our next webinar which is on the 30th of May which is lessons learned from the Karelian claps park for a sore third we've done one which were available on the YouTube channel about assessing suppliers so that you don't select a supplier who is in financial issues we've done one on how to protect yourself through the contract although by the time you sign the contract and the contractors looking insolvent it's a reactive if you like you're minimizing the impact of the risk rather than likelihood and this one is by Alistair who's a mega project manager and he was managing and projects internationally where two of his supplies started the descent into insolvency so what are the actual signs on the ground how do you manage that process as they are in the downward death spiral and to some extent how do you actually manage it when they do actually become insolvent and cease to exist there's also a traditional one procurement for benefits and value workshop so the challenges there is I want to only oh and I'm doing a project cause I want benefits and value from it I want a supplier who does that but actually often what they deliver the tangibles are a bit remote or there's others involved in actually delivering the benefit and value and outside their control so how did you select people had you contract on that basis having it's a sort of workshop degree approach and having looked on the website this morning it's actually fully so with that I'd like to say a big thank you to Martin and others I'll say the others first of all David Wally who's done a lot of back-end stuff which you hadn't seen and named this webinar to work efficiently particularly with men Tomita and of course you for participating and above all Martin for actually contributing so with that thank you very much good bye

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *