Apr 07, 2020 / by Dave Wells Data Management

Data Stewards as Problem Solvers: A Practical Tool for the Job

A Tool for Data Stewards to Fix Common Data Problems

Data stewardship is an essential role in modern data governance. Every data-driven organization needs to have stewards who can quickly resolve data management problems and challenges. Data stewards facilitate consensus about data definition, quality, and usage. They guide activities to complete metadata, improve data quality, and ensure regulatory compliance. Stewards are also responsible to make recommendations about data access, security, distribution, retention, archiving, and disposal.

Unfortunately, typical data stewardship practices often don’t measure up to the importance of the role. All too frequently, data stewards are identified and assigned responsibility without the time and training to do the job well. When we designate busy people as data stewards without making time for them to do stewardship work we should not expect high-impact results. Nor should we expect success without training stewards about roles, relationships, and accountabilities related to data.

Along with time and training, data stewards need tools that help them to do their work. This article offers a simple tool to help diagnose data problems, find the path from symptoms to causes, and then get from causes to solutions. The tables below identify several common symptoms of data challenges that data stewards frequently encounter, grouped by ten core data management processes – naming data, defining data, designing data, managing quality, integrating data, accessing data, managing metadata, administering databases, managing systems, and governing data. Common causes of and solutions to data problems are identified for each process.

To use the tool, begin by browsing the index of common symptoms to find those related to your data management issues. Then use the associated numbers to find each symptom in the process tables. Note that a single symptom is often listed in several of the process tables. Explore the processes, causes, and solutions to develop problem-solving ideas and plans.

Index: Common Symptoms of Data Problems

application integration difficulty integration difficulty	47	inefficient business analysis	26, 39
business rule violations in data	31, 40, 90	insufficient data storage capacity	72
can’t access needed data	52	lack of data definitions	10
complex system interfaces	48, 81	lack of trust in data	32
conflicting documentation	64	large change request backlog	20
confusing abbreviations	5	limited data sharing	49
confusing documentation	67	lost data can’t be recovered	80
corrupted data can’t be repaired	79	meaningless data definitions	12
data consolidation difficulties	98	meaningless data names	1
data not available when needed	58	missing documentation	62
data ownership conflicts	99	misunderstood data	15, 69
data privacy compromised	53, 55, 89, 93	multiple names & aliases	6
data retention/disposal uncertainty	97	need for data standardization	100
data security compromised	54, 88, 92	needed access not authorized	56
data-related compliance violations	94	needed features not implemented	78
difficult-to-use data	28, 37	non-unique data names	2
disaster recovery uncertainties	95	obsolete data definitions	13
enterprise reporting difficulty	46	obsolete permissions still active	57
excessive database downtime	76	outdated documentation	66
failure to meet business needs	21	overlapping and conflicting data	44
hard to find data definitions	14	poor application performance	83
hard to find documentation	65	poor data access performance	60
hard to find needed data	51	poor data quality	86, 91
hard-to-navigate databases	29, 59	poor database performance	30
hard-to-identify data	8	poor query performance	74
hard-to-navigate databases	29	poor structural integrity	19, 33
high level of data disparity	9, 17, 25, 43, 70, 85	poor update performance	75
high level of data redundancy	18, 27, 71, 84	shadow databases	41
inadequate metadata	87	shadow systems & databases	23
inappropriate use of data	16	spreadsheet proliferation	22, 42, 50, 61
incomplete data	36	structureless data names	4
incomplete documentation	63	territorialism inhibits data sharing	96
incorrect data	34	unanticipated growth problems	73
incorrect data definitions	11	unnamed data components	7
incorrect data names	3	unreliable database connections	77
incorrect reporting	24, 38

Naming Data

	Symptoms	Causes	Solutions
1	meaningless data names	informal naming practices lack of naming standards standards	data naming taxonomy data naming vocabulary standard naming structure standard abbreviations list compliance incentives
2	non-unique data names
3	incorrect data names
4	structureless data names
5	confusing abbreviations
6	multiple names & aliases
7	unnamed data components
8	hard-to-identify data
9	high level of data disparity

Defining Data

	Symptoms	Causes	Solutions
10	lack of data definitions	lack of data definition standards poor data definition practices lack of business participation legacy databases disparate metadata	data definition standards data definition templates data definition wiki business/tech collaboration data definition review metadata repository definitions system-of-record
11	incorrect data definitions
12	meaningless data definitions
13	obsolete data definitions
14	hard to find data definitions
15	misunderstood data
16	inappropriate use of data
17	high level of data disparity
18	high level of data redundancy

Designing Data

	Symptoms	Causes	Solutions
19	poor structural integrity	poor modeling techniques wrong choice of model type poor business representation excessive detail insufficient detail process-oriented design application-oriented design	data model standards E-R model guidelines dimensional model guidelines normalization guidelines atomic data guidelines aggregate data guidelines subject-oriented design consumer-oriented design
20	large change request backlog
21	failure to meet business needs
22	spreadsheet proliferation
23	shadow systems & databases
24	incorrect reporting
25	high level of data disparity
26	inefficient business analysis
27	high level of data redundancy
28	difficult-to-use data
29	hard-to-navigate databases
30	poor database performance
31	business rule violations in data

Managing Data Quality

	Symptoms	Causes	Solutions
32	lack of trust in data	poorly defined DQ rules missing DQ rules absence of quality measures absence of quality reporting lack of accountability incomplete/incorrect edits	DQ rules taxonomy defined DQ rules DQ metrics and measures published DQ reports regular DQ audits designated DQ accountability DQ tasks in project plans
33	poor structural integrity
34	incorrect data
35	untimely data
36	incomplete data
37	difficult-to-use data
38	incorrect reporting
39	inefficient business analysis
40	business rule violations in data
41	shadow databases
42	spreadsheet proliferation

Integrating Data

	Symptoms	Causes	Solutions
43	high level of data disparity	lack of integration architecture technology-driven integration inadequate data warehouse absence of data marts unmanaged master data poor integration practices missing/wrong data sources	sound integration architecture business-driven integration sound warehousing design targeted data marts master data management integration best practices defined data sourcing criteria
44	overlapping and conflicting data
45	untraceable data
46	enterprise reporting difficulty
47	application integration difficulty
48	complex system interfaces
49	limited data sharing
50	spreadsheet proliferation

Accessing Data

	Symptoms	Causes	Solutions
51	hard to find needed data	missing metadata inadequate data access tools insufficient indexing inadequate search capability lack of content management poor user interface excessive downtime database design not user friendly ineffective performance tuning ineffective security processes	robust metadata user-friendly tools and interfaces indexing and searching data access portals service level agreements service level accountability published service level metrics security policies & procedures periodic security/privacy audits security/privacy accountability
52	can’t access needed data
53	data privacy compromised
54	data security compromised
55	data privacy compromised
56	needed access not authorized
57	obsolete permissions still active
58	data not available when needed
59	hard-to-navigate databases
60	poor data access performance
61	spreadsheet proliferation

Managing Metadata

	Symptoms	Causes	Solutions
62	missing documentation	casual metadata management fragmented metadata tools lack of documentation standards lack of data modeling standards undocumented changes no documentation incentives no documentation reviews “rush to production” projects	metadata templates & guidelines project metadata standards maintenance metadata standards metadata registries & portals metadata system-of-record metadata accountability metadata tasks in project plans incentives and reviews
63	incomplete documentation
64	conflicting documentation
65	hard to find documentation
66	outdated documentation
67	confusing documentation
68	untraceable data
69	misunderstood data
70	high level of data disparity
71	high level of data redundancy

Administering Databases

	Symptoms	Causes	Solutions
72	insufficient data storage capacity	ineffective storage management passive growth management ineffective performance tuning unscheduled maintenance inadequate database connectivity outdated DBMS versions insufficient backup & recovery	continuous capacity planning proactive growth management performance SLAs availability/uptime SLAs connection protocol standards connectivity SLAs routine DBMS upgrades backup & recovery practices
73	unanticipated growth problems
74	poor query performance
75	poor update performance
76	excessive database downtime
77	unreliable database connections
78	needed features not implemented
79	corrupted data can’t be repaired
80	lost data can’t be recovered

Managing Systems

	Symptoms	Causes	Solutions
81	complex system interfaces	lack of data sharing architecture lack of integration architecture poor application design “quick fix” maintenance “misfit” acquired systems inconsistent data formats limited reuse of data functions testing with production data	application architecture standards application design review maintenance & testing standards application acquisition guidelines data sharing incentives database wrappers SOA-based access & update reusable data quality rules managed test data
82	untraceable data
83	poor application performance
84	high level of data redundancy
85	high level of data disparity
86	poor data quality
87	inadequate metadata
88	data security compromised
89	data privacy compromised
90	business rule violations in data

Governing Data

	Symptoms	Causes	Solutions
91	poor data quality	lack of data management goals unclear, uncertain, ambiguous or misaligned RAA (responsibility, authority, & accountability) poor P&P (policies & procedures) understaffed data management underfunded data management	"data as an asset" culture clear data management goals quality RAA + P&P security RAA + P&P privacy RAA + P&P compliance RAA + P&P disaster recovery RAA + P&P designated data ownership
92	data security compromised
93	data privacy compromised
94	data-related compliance violations
95	disaster recovery uncertainties
96	territorialism inhibits data sharing
97	data retention/disposal uncertainty
98	data consolidation difficulties
99	data ownership conflicts
100	need for data standardization

Ultimately, data stewards are problem solvers. In the best of circumstances they can intervene early in data management processes to prevent problems, but realistically they spend much of their time seeking a resolution to problems that already exist. Effective problem solving for data issues depends on diagnosis, causal analysis, and solution planning. This tool is designed with those goals in mind, giving structure to the path from symptoms to causes and from causes to solutions.

Previous post by expert Next post by expert

Dave Wells

Dave Wells is an advisory consultant, educator, and industry analyst dedicated to building meaningful connections throughout the path from data to business value. He works at the intersection of information...

More About Dave Wells